Understanding Economic Policy Reform
Journal of Economic Literature, Vol. XXXIV (March 1996), pp. 9–41
Rodrik: Understanding ofEconomic Policy Reform Journal Economic Literature, Vol. XXXIV (March 1996)
Understanding Economic Policy Reform
By DANI RODRIK Columbia University
This paper has benefited from the comments of Herschel Grossman, Anne Krueger, Gustav Ranis, Jeffrey Sachs, John Williamson, and three referees.
I. Introduction
HIS GOVERNMENT will be austere, uncompromising, and unpopular if that is what is required to achieve economic recovery,” declared Mario Soares in 1983 upon taking office as prime minister of Portugal.1 This kind of talk is music to the ears of economists. Reform requires austere policies which respect budget constraints. It also precludes compromising with the narrow, special interest groups that have been the beneficiaries of the deleterious policies of the past. Policy makers who have courageously taken on such interest groups and pursued market-oriented policies are the heroes of the economics profession (see Arnold C. Harberger 1993). That such policies should also be unpopular (as Soares feared they would), however, requires a lot more explaining. What is the point of loudly proclaiming reforms if these are not aimed at improving the well-being of a large majority of the population? And if that is their goal, why should reforms be unpopular? In many areas of policy, there may ex1 Quoted by José Maria Maravall in Luiz Carlos Bresser Pereira, Maravall, and Adam Przeworski (1993).
“
T
ist “technical” uncertainty as to what the appropriate solution is to the problems at hand. Think of President Clinton’s health care plan, for example, or of global warming. Consequently, reforms will arouse opposition if they are viewed as applying the wrong fix or if they are perceived as being primarily redistributive (that is, zero-sum). What is remarkable about current fashions in economic development policy (as applied to both developing and transitional economies), however, is the extent of convergence that has developed on the broad outlines of what constitutes an appropriate economic strategy. This strategy emphasizes fiscal rectitude, competitive exchange rates, free trade, privatization, undistorted market prices, and limited intervention (save for encouraging exports, education, and infrastructure). Faith in the desirability and efficacy of these policies unites the vast majority of professional economists in the developed world who are concerned with issues of development.2
2 The convergence is not complete of course. But compared to two decades ago, the various sides have moved substantially closer to each other. One indicator of this is the recent book by Bresser Pereira, Maravall, and Przeworski (1993), which advocates a “social democratic” approach. The views expressed in this book concede an inor-
9
10
Journal of Economic Literature, Vol. XXXIV (March 1996) we observe such instances of collective irrationality. The events of the last decade have underscored the need to understand the political-economy of policy making. One of the eventual consequences of the global debt crisis that erupted in 1982 was a wave of market-oriented economic reforms, the likes of which have never been seen. The reforms were strongest and most sustained in Latin America, where countries like Bolivia, Mexico, Argentina, Peru, Colombia, and Brazil joined Chile in orthodoxy. But this was very much a global phenomenon. “Stabilization” and “structural adjustment” became the primary preoccupation of government leaders in Asia and Africa as well, even though the commitment to economic orthodoxy varied across countries and over time. These countries were in turn soon joined by the previously socialist economies of Eastern Europe and the former Soviet Union. Economists who had cut their teeth in Latin America’s economic quagmires became the advisors and analysts of these transitional economies. Even India, the giant archetype of a closed, import-substituting economy among developing countries, embarked on a process of economic liberalization in 1991 (see Jagdish Bhagwati 1993 and Arvind Panagariya 1994). These reforms were encouraging to economists and a vindication of sorts to those among them who had long advocated market-oriented reforms. But they in turn raise their own puzzles. Most fundamental of all, why are so many governments reforming now, after decades of adherence to policies of an opposite kind? This question poses a particularly important challenge to political economists: an understanding of these countries’ experiences now requires a theory that explains not only why seemingly dysfunctional policies had been initially un-
Hence economists are often torn between two conflicting perspectives: on the one hand, good economic policy should produce favorable outcomes and therefore should prove also to be good politics; on the other hand, the implementation of good economic policy is often viewed as requiring “strong” and “autonomous” (not to say authoritarian) leadership. The experience of Chile, a country which has perhaps gone further than any other in implementing liberal economic policies, provides a good example. An essay on Chile’s reform strategy by José Piñera (1994), an economist and minister of labor and social security under General Pinochet, concludes: “[i]n the end, good policy is good politics” (p. 231). The irony is that most of the reforms the author glowingly discusses in the preceding pages required the suspension of normal politics and as heavy a dose of authoritarianism as seen anywhere. Good economics does often turn out to be good politics, but only eventually. Policies that work do become popular, but the time lag can be long enough for the relationship not to be exploitable by would-be reformers. In Chile’s case, free market policies (implemented after 1973) were eventually resoundingly endorsed in the presidential elections of 1989 and have become the envy of Latin America.3 Conversely, bad economics can be popular, if only temporarily. President Alan García’s popularity soared in Peru during his first two years in office (1985–86), thanks to expansionary fiscal policies whose medium-term unsustainability should have been obvious to anyone with common sense (see Ricardo Lago 1991). The puzzle is why dinate amount to the consensus view, and depart from it in remarkably few details. I will discuss this book in Section IV. 3 For a recent evaluation, see Barry Bosworth, Rudiger Dornbusch, and Ral Labn (1994).
Rodrik: Understanding Economic Policy Reform dertaken and then maintained for so long, but also why these policies were suddenly abandoned en masse during the 1980s, often by the same politicians who had been among their most ardent supporters. Second, while the reforms were inspired at least in part by the East Asian experience, they took place much more quickly and, in many areas, are going considerably beyond those undertaken in East Asia. This raises the question of whether the new wave of reformers have internalized the correct lessons from the East Asian experience. Finally, are there any helpful rules for reformers to follow in guiding their policies through complicated political terrain? Can one hope to develop a “how-to” manual for the reformist politician? Puzzlement over such questions has led to a large and growing literature. A very short bibliography would include books by Merilee Grindle and John Thomas (1991), Robert Bates and Krueger (1993), Krueger (1993), Przeworski (1991), Ranis and Syed Mahmood (1992), Bresser Pereira, Maravall, and Przeworski (1993), Stephan Haggard and Robert Kaufman (1992), Dornbusch and Sebastian Edwards (1993), Haggard and Steven Webb (1994), Lance Taylor (1994), Williamson (1994), and Ian Little et al. (1993), not to mention countless papers. 4 As this partial list indicates, both economists and political scientists have devoted their attention to these issues, often together in coauthored or coedited works. Indeed, no other area of economics or political science that I can think of has spawned so much interdisciplinary work.5 In this essay, I will provide an econo4 One recent survey—Mariano Tommasi and Andres Velasco 1995—which overlaps with this one deserves special mention. 5 The literature on the economics of policy reform is of course even larger. For recent surveys, see Vittorio Corbo and Stanley Fischer (1995) and Rodrik (1995b).
11
mist’s perspective on the political economy of policy reform. I begin by examining the origins and analytical content of the new orthodoxy in development policy (Section II). I will focus here on two issues in particular which I feel remain in need of clarification. One of these concerns the distinction between (a) macroeconomic policies aimed at economic stability, such as fiscal, monetary, and exchange rate policies, and (b) liberalization policies aimed at structural reform and growth, such as the removal of relative-price distortions and the reduction of state intervention. It has become commonplace to conflate these two groups of policies, but for analytical purposes they are best kept apart. As we shall see, they also have different political-economy underpinnings. Moreover, maintaining the distinction reminds us that the consensus on what constitutes appropriate structural reform is based on much shakier theoretical and empirical grounds than is the consensus on the need for macroeconomic stability. The second issue concerns the appropriate lessons to be drawn from the experience of East Asian success stories. The new orthodoxy has tended to draw a somewhat biased picture that needs correction. Next, I will turn to the reforms of the 1980s and 1990s. This experience has opened an important window on the motivations of politicians, as well as on the nature of interactions between the economy and the polity. As indicated above, an important question is why so many countries have suddenly caught the reform bug. The confluence of economic crisis with reform has led to the natural supposition that crisis is the instigator of reform, a hypothesis that keeps reappearing in the literature and yet is inadequately analyzed. Section III discusses this issue, as well as related questions such as: what, if any, are the short-term
12
Journal of Economic Literature, Vol. XXXIV (March 1996) parently successful example of Soviet planning invited emulation. Moreover, the economic ideas of the 1950s and early 1960s tended to dismiss the benefits from trade and emphasized the need for physical capital accumulation and infant-industry promotion. As Krueger puts it,
[t]he underlying premises regarding markets and governments implicit in these policy prescriptions are obvious: There was a strong emphasis on the primacy of market imperfections. Market failures were thought to be relatively strong, while it was assumed that governments could correctly identify and perform economic functions. Virtually no attention was given to the possibility that there might be government failure. (p. 49)
costs of reform? and does foreign aid help or hamper reform? Finally, I will discuss some normative aspects related to the politics of policy reform (Section IV). In particular, I will analyze what kind of guidance, if any, the literature provides to the reformist politician regarding the appropriate strategies for carrying out his program. The focus here will be on the conflict between top-down versus participatory approaches. This discussion will take us back to some of the deeper questions about the collective rationality of political institutions and behavior. II. The New Orthodoxy in Development Thinking and Enduring Puzzles Once upon a time, there was something called the Third World and we thought we understood how it worked (or did not work). Countries of the Third World followed import-substitution policies, so called because the overarching objective of economic policy was to develop domestic manufacturing capability for goods previously imported. Such policies included import controls, overvalued exchange rates, binding ceilings on interest rates, a heavy dose of public ownership, and pervasive price regulation. The political-economy counterpart of these policies was the predominance of urban over rural interests, and within the urban sector, an uneasy alliance of sorts between the protected industries and the bureaucrats administering the protection. Perhaps no other economist has done as much as Anne Krueger in documenting and popularizing the shortcomings of import-substituting policies. As she points out in her Ohlin Lectures (1993), there was a multitude of reasons for the initial adoption of these policies. Newly independent governments had a strong desire for industrialization, and the ap-
And there was plenty of government failure. Not only did many infant industries fail to mature, but many countries succumbed to stop-go cycles driven by excessive government spending. Krueger explains that in viewing the government as a “benevolent social guardian” most economists had ignored a number of important forces at work. Individuals in the public sector were apt to follow their own selfish interests. They would be lobbied by pressure groups aiming to impose their own agenda on a largely docile majority. Policy interventions would create rent-seeking incentives diverting entrepreneurs from productive activities. Finally, the informational disadvantage of government bureaucracies over market participants would doom even the best-laid plans to inefficiency. Meanwhile, four East Asian economies were making a mockery of the export pessimism that had persuaded policy makers elsewhere to follow an inwardoriented strategy. South Korea and Taiwan, in particular, were able to engineer a remarkable increase in their growth rates, thanks to sharp jumps in their investment and export efforts during the
Rodrik: Understanding Economic Policy Reform
TABLE 1 COMPARATIVE GROWTH EXPERIENCE Per capita GDP, 1960 (1985 dollars) 883 1359 873 1017 1128 1745 2798 3294 Per capita GDP, 1989 (1985 dollars) 6206 8207 815 1082 756 4138 5163 3608
13
Country South Korea Taiwan Ghana Senegal Mozambique Brazil Mexico Argentina
Per capita GDP growth, 1960–89 (%) 6.82 6.17 −0.54 0.16 −2.29 3.58 2.36 0.63
Source: Penn World Table 5.5 (1993).
mid-1960s. Table 1 summarizes this impressive performance in comparative context. Recent work by Jong-Il Kim and Lawrence J. Lau (1992) and Alwyn Young (1993, 1994) has shown that these were miracles of accumulation rather than of productivity: the sharp increases in physical and human capital as well as in labor-force participation account for virtually all of the rise in output, and consequently the East Asian tigers’ performance with respect to total factor productivity (TFP) growth does not look outstanding. Of course, for accumulation to have taken place at such rates, the profitability of investment in the region must have been very high, which needs explanation. What was the key to these economies’ success? Among professional economists, there soon developed the view that the East Asian miracles could be attributed to market-oriented policies and the reduced role of government intervention. Hence, Ranis and Mahmood (1992, p. 138) attribute the Taiwanese and South Korean successes to the willingness of the governments in both countries—albeit somewhat less so in South Korea—to allow growth to proceed along a “natural” path and an aversion to use covert
measures of resource transfers in order to promote growth artificially.
And Krueger (1993, p. 30) writes:
To be sure, the Korean economy has not been characterized and is not characterized by laissez-faire. But in contrast to the overcontrolled, overregulated, highly distorted economies described above, the Korean economy has been characterized by diminishing intervention in most spheres of economic activity, and the degree of distortion is considerably smaller.
Indeed, in both Korea and Taiwan, there was an extensive set of reforms during the late 1950s and early 1960s, on which more later. So the new orthodoxy was built on two mutually reinforcing pillars: one was the set of policies that had been tried by the import-substituting countries and had failed; the second was the set of successful policies implemented by the East Asian tigers. These two sets of policies bear close scrutiny, as they overlap to a much greater extent than the orthodox case likes to admit. A. Macroeconomic Disequilibrium or Import-Substitution Policies? In describing the experience of developing countries it has become common
14
Journal of Economic Literature, Vol. XXXIV (March 1996) rates are different. By definition, they result in trade deficits that are unsustainably large, and therefore in balanceof-payments crises (unless they are corrected quickly). Consequently, exchange rate misalignments are closely associated with economic instability and with the deterioration of economic performance over the medium to long run. The confusion between these two types of policies often reveals itself in empirical studies that uncover a negative relationship between measures of exchangerate distortion and economic growth, and then attribute the effect to the lack of openness in the sense of trade protection. 7 Similarly, consider subsidies to specific industries versus large fiscal deficits. The former can be damaging if the targeted industries do not produce positive externalities or if the subsidies lead to rent-seeking activities. But the damage done by large, sustained budget deficits is often much larger. An unsustainable deficit can be financed by foreign borrowing or by monetization, and in either case a crisis is often not too far away. 8 Usually, external sources are the first resort, and an inflation problem follows on the heels of a payments crisis. The World Bank recently undertook a large-scale research project on the consequences of public-sector deficits in developing countries. The findings were unequivocal: “Deficits . . . are unambiguously bad for growth.”9 This distinc7 See for example the use made of David Dollar’s (1992) “real exchange rate distortion” index in World Bank (1993). 8 Of course, foreign borrowing or moderate amounts of monetization (seignorage or inflation tax) need not be bad things. The focus here is on unsustainable levels of borrowing or seignorage. 9 William Easterly, Carlos Alfredo Rodrguez, and Klaus Schmidt-Hebbel (1994, p. 1). It is of course possible, as a referee pointed out, that a deeper underlying cause (i.e., political instability) is jointly responsible for both the deficits and low growth.
to lump together a wide range of policies, as I too have done above, under the label of “import-substitution policies.” For descriptive purposes, this makes perfect sense. Except for a handful of countries in East Asia, most developing countries did combine illiberal trade and price policies with (at least occasional) fiscal profligacy and overvalued exchange rates. So there was a common syndrome, which perhaps does deserve to go under a single name. However, the practice has also frequently led economic analysis astray and generated confusion. The trouble is that failures were often misattributed to microeconomic policies, when their sources lay either with unsustainable macroeconomic policies or bureaucratic and institutional shortcomings. Consider for example two trade-related policies: import restrictions (e.g., import quotas) and overvalued exchange rates. These two have rather different implications for economic stability and long-run performance. The effect of import restrictions is to repress trade (imports as well as exports in the longer run). While this is costly insofar as it engenders some resource misallocation, it does not inherently generate economic instability, nor does it necessarily reduce long-term growth.6 Overvalued exchange
6 In traditional economic theory, trade restrictions have level effects, but no growth effects. That is, a 20 percent tariff may reduce real income by, say, 0.5 percent of GDP (permanently), but it will not affect the economy’s long run growth rate. In the more recent endogenous growth literature, trade restrictions may have growth effects, but these growth effects can go either way depending on the model. For a small economy, the effect tends to depend on whether the comparativeadvantage effect pulls resources in or out of the growth-sectors of the economy (those with technological externalities or learning-by-doing). Empirical work on growth has often claimed to discover a negative relationship between trade protection and growth. However, this literature is marred by severe analytical and conceptual confusions, with macroeconomic policies often confused for trade restrictions and the endogeneity of trade policy not fully accounted for.
Rodrik: Understanding Economic Policy Reform
TABLE 2 DETERMINANTS OF THE DEBT CRISIS, 1982 Failure to adjust monetary and fiscal policy Yes Yes Yes Yes Yes Yes Yes Yes
15
Large external shock Troubled Counties Argentina Brazil Chile Costa Rica Cote d’lvoire Mexico Morocco Nigeria Unweighted Average Moderately Troubled Countries Colombia Kenya Sri Lanka Unweighted Average Untroubled Countries Cameroon India Indonesia Korea Pakistan Thailand Turkey Unweighted Average No Yes Yes No Yes No No No
Index of relative-price distortion 0.3054 0.2019 0.4460 0.2818 0.2438 n.a. 0.2675 0.2306 0.2824 0.2744 0.1218 0.8606 0.4189 0.2344 0.2620 0.4503 0.2128 0.3814 n.a. n.a. 0.3082
No Yes Yes
Yes Yes No
Yes No No Yes No Yes No
No No No No No No No
Source: Little et al. (1993), Table 4.4, except for the relative-price distortion index which is taken from Easterly (1993). The latter index is the variance of the log input prices (relative to U.S. prices) across commodities, measured in 1980. See Easterly (1993) for the method of calculation and the justification for the index.
tion—between micro and macro—is lost in many studies that attribute balanceof-payments crises to activist industrial policies (rather than the unwillingness or inability to balance budgets). The watershed event of the 1980s for most developing countries, the generalized debt crisis that followed the Mexican moratorium of August 1982, was a dramatic confirmation of the importance of prudent macroeconomic policies. Looking below the surface, it was evident that the crisis affected only those countries that did not respect budget constraints. India, the import-substitut-
ing country par excellence, managed to escape the debt crisis during the 1980s, thanks to its tradition of conservative monetary and fiscal policies.10 At the other end of the spectrum, South Korea experienced a payments crisis in 1979– 80, before the Latin American countries, as a consequence of an ambitious investment program running ahead of available domestic savings. But the quick adoption
10 During the 1980s, however, a growing publicsector deficit began to endanger macroeconomic balances. In 1990, the deficit reached 8.3 percent of GDP, culminating in a balance-of-payments crisis in 1991 (Panagariya 1994, pp. 205–06).
16
Journal of Economic Literature, Vol. XXXIV (March 1996) most everything right on the microeconomic front, but still face a deep crisis if macro policies—in Mexico’s case an overvalued exchange rate aggravated by pre-election credit expansion—are not managed well. This distinction between microeconomic distortions and macroeconomic stability is one that economists have long recognized. Yet it is also one that has made little impression on the development profession. In an essay published in 1975, for example, Carlos Díaz-Alejandro (1975) faulted the Little-Scitovsky-Scott (1970) study mentioned above for lumping together all features of the import-substitution sydrome, such as import and other controls, tariffs, overvalued and pegged exchange rates, spectacular balance-of-payments crises, inflationary pressures, and stop-go cycles. Following a “guilt by association” procedure, they then tend to blame much of what is going wrong in less developed countries on that ill-defined syndrome. Unsophisticated readers may indeed conclude that nearly everything gone wrong in those countries is due to that wicked syndrome. Consider a mental experiment. What would have happened if, say, Argentina and Colombia had adopted flexible exchange rates back in 1945, while adopting also an across-theboard import tariff of 150 per cent ad valorem? I suspect their record, at least on growth and exports, would have been much better. Their harmful stop-go policies may be blamed to a large extent on exchange-rate mismanagement . . . and on other short-run policies that could be analytically separated from the long-run effects of protection. (Díaz-Alejandro 1975, pp. 115–16, emphasis in the original)
of expenditure-reducing and expenditure-switching policies enabled a swift recovery, after a five percent contraction in 1980 (see Susan Collins and Won-Am Park 1989). Before long, Korea’s crisis was forgotten. More broadly, there was no correlation between the propensity to fall into crisis and the nature of microeconomic policies proper. As Table 2 indicates, the countries that experienced a debt crisis in 1982 were those that failed to adjust their monetary and fiscal policies, and not those that had large microeconomic distortions or were confronted with particularly large external shocks. There is little question that the microeconomic distortions in themselves were often costly. A series of team efforts led by Little, Tibor Scitovsky, and Maurice Scott (1970), Bela Balassa (1971), Bhagwati (1978), and Krueger (1978) had documented in detail the resource-allocation costs of high protection. These studies used measures of effective rates of protection (ERP) and domestic resource costs (DRC) to demonstrate that the governments’ price and trade policies had led to the creation of wildly inefficient industries, at times producing negative value-added at world prices! Moreover, the net incentive effects of these policies seemed haphazard, judged by the variability of ERPs and DRCs across industries. More recent work has shown that particular relative price distortions, those affecting capital and intermediate goods, can be costly to growth as well (Bradford De Long and Lawrence Summers 1991; Easterly 1993). However, what eventually drove many import-substituting countries to ruin were not such micro- economic inefficiencies, but macroeconomic imbalances and the inability to correct them with sufficient speed. Mexico’s currrency crisis in December 1994–January 1995 is a poignant example of how one can get
Elementary as these points may be, they were largely overlooked in the aftermath of the debt crisis. The consensus post mortem view held the whole complex of import-substitution policies responsible for what was essentially a crisis of overspending exacerbated by the fickleness of international capital markets. It became commonplace to view the debt
Rodrik: Understanding Economic Policy Reform
TABLE 3 THE “WASHINGTON CONSENSUS” AND EAST ASIA Elements of the Washington Consensus 1. Fiscal discipline 2. Redirection of public expenditure priorities towards health, education and infrastructure 3. Tax reform, including the broadening of the tax base and cutting marginal tax rates 4. Unified and competitive exchange rates 5. Secure property rights
17
South Korea Yes, generally Yes Yes Yes
Taiwan
Yes, generally
Yes
6. Deregulation 7. Trade liberalization 8. Privatization
Yes (except for limited time periods) President Park starts his rule in 1961 by imprisoning leading businessmen and threatening confiscation of their assets. Limited Limited until the 1980s No. Government established many public enterprises during 1950s and 1960s. DFI heavily restricted Limited until the 1980s
Yes Yes
9. Elimination of barriers to direct foreign investment (DFI) 10. Financial liberalization.
Limited Limited until the 1980s No. Government established many public enterprises during 1950s and 1960s DFI subject to government control Limited until the 1980s
Source: Williamson (1994) for first column, and author’s evaluation.
crisis as the consequence of import-substitution (“inward-oriented”) policies. 11 The intellectual ground was therefore cleared for the wholesale reform of prevailing policies in Latin America, Africa, and Asia. Orthodox economists who had the ear of policy makers now had their chance to wipe the slate clean and mount a frontal attack on the entire range of policies in use. After some de11 This view was (and is) strongly held in the World Bank, which shaped development policy in many countries during the 1980s through its structural adjustment loans. See for example the World Bank volume edited by Vinod Thomas et al. (1991), and especially the essays by the editors and by Ernest Stern.
lay, this produced dramatic results, especially in Latin America. B. What Did the East Asian Governments Do Right? Ironically, many governments (notably in Latin America) ended up implementing policies that went far beyond what the East Asian governments themselves had adopted since the 1960s. As intimated above, many of the discredited policies had long been in use in South Korea, Taiwan, and Singapore, and apparently to good effect. (Hong Kong’s policies have come closest to the laissezfaire ideal, even though there is plenty of
18
Journal of Economic Literature, Vol. XXXIV (March 1996) fered in the investment decisions of private enterprises. And far from privatizing public enterprises, both countries actually increased their reliance on such enterprises during the crucial decade of the 1960s. 13 In short, where South Korea and Taiwan followed the orthodox path most closely was in maintaining conservative fiscal policies and competitive exchange rates; this accounts for their ability to avoid protracted periods of macroeconomic instability, particularly in the crisis-ridden decade of the 1980s. In the area of microeconomic interventions, however, their experience diverged from the orthodox path. By contrast, it is striking how many Latin American countries have come within reaching distance of completing the items on the “Washington consensus” in a period of no more than a few years during the 1980s. Mexico, Bolivia, and Argentina, to cite some of the more distinguished examples, have undertaken more trade and financial liberalization and privatization within five years than the East Asian countries have managed in three decades. I will discuss the reforms of the 1980s at greater length later. First we have to confront two puzzles that arise from the contrasting experiences of the East Asian tigers and other developing countries prior to the 1980s. How could the East Asian countries avoid the disasters that accompanied interventionist policies elsewhere? And why did most developing countries succumb so easily, and often periodically, to unsustainable fiscal and exchange-rate policies? Neither of the puzzles is handled very well in the existing literature, although we do have a number of useful leads.
13 Probably the two best known works on these issues are Alice Amsden (1989) and Robert Wade (1990). Leroy Jones and Il Sakong (1980) and Ching-yuan Lin (1973) are two older books which are extremely detailed and informative.
intervention in the housing market.) To appreciate this point, it is useful to spell out in somewhat greater detail the elements of the new orthodoxy, so we can compare them to East Asian policies. Fortunately, we have a useful list of policy desiderata compiled by Williamson, who has dubbed it the “Washington consensus.” Williamson originally compiled the list in 1990, and a summary is included in Williamson (1994). In Williamson’s words, “[t]he ‘Washington consensus’ offers a description of what is agreed about the set of measures that are typically called for in the first stage of policy reform. . . .” (p. 17). 12 The list is shown in Table 3, along with my own summary comments on the degree of compliance exhibited by Taiwan and South Korea in each area of reform. How well did South Korea and Taiwan do according to this list? Judging by the number of the prescriptions these countries did or did not follow, we would have to award South Korea a score of about five (out of ten), and Taiwan about six. Both countries managed fiscal expenditures and revenues rather well, avoiding macroeconomic stop-go cycles and high inflation. They also consistently maintained unified exchange rates (Taiwan since 1961 and Korea since 1964) and competitive parities for the most part. But the rest of the scorecard is less spectacular. Taiwan welcomed DFI, but Korea much less so. Korea repressed interest rates and made heavy use of subsidized credit; Taiwan did not do so, but did give priority to public enterprises in credit allocation. Neither country significantly liberalized its import regime until the 1980s. Both countries heavily inter12 Williamson is quick to say that there remains a number of disagreements among economists on some of the narrower issues, for example the speed of trade liberalization or the need to eliminate indexation. See pages 17–18 for his list of disagreements.
Rodrik: Understanding Economic Policy Reform
TABLE 4 HUMAN CAPITAL INDICATORS IN EAST ASIA: ACTUAL VERSUS PREDICTED VALUES FOR THE EARLY 1960s Primary enrollment ratio, 1960 Hong Kong Indonesia Japan Korea Malaysia Singapore Taiwan Thailand Act. Pred. Diff 0.87 0.83 0.04 0.67 0.51 0.16 1.03 0.92 0.11 0.94 0.57 0.37 0.96 0.68 0.28 1.11 0.78 0.33 0.96 0.62 0.34 0.83 0.57 0.26 Secondary enrollment ratio, 1960 Act. Pred. Diff 0.24 0.23 0.01 0.06 0.07 −0.01 0.74 0.29 0.45 0.27 0.10 0.17 0.19 0.15 0.04 0.32 0.21 0.11 0.28 0.12 0.15 0.12 0.10 0.02
19
Literacy rate, 1960 Act. 0.70 0.39 0.98 0.71 0.53 0.50 0.54 0.68 Pred. Diff 0.59 0.11 0.25 0.15 0.70 0.28 0.31 0.40 0.43 0.10 0.54 −0.04 0.36 0.18 0.31 0.37
Source: Authors’ calculations. Note: Predicted values of the indicators are obtained from a cross-country regression run on a 118-country sample, with per capita GDP in 1960 and its square used as independent variables. Source of the data: Alan Heston and Robert Summers (1988) and Robert Barro and Hoger Wolf (1989).
C. How Did East Asian Countries Manage to Intervene without Inviting Rent Seeking? One major puzzle noted above had to do with the apparently successful pursuit of interventionist microeconomic policies by South Korea and Taiwan. Why did trade protection, industrial policy, and subsidized credit work in these countries when it failed most everywhere else? There has been much debate about these questions, not the least about the degree of intervention itself. The extent to which activist government policies enabled these miracles of accumulation to take place is an issue that is not yet settled. 14 From our present perspective, we can ignore purely economic aspects and narrow down the question to the following: how were the East Asian governments able to avoid the rent-seeking activities that typically accompanied microeconomic interventions? While we have some clues, the answer is that we do not really know. There was
14 For two recent contrasting viewpoints, see World Bank (1993) and Albert Fishlow and others (1994).
clearly something special about the ability of the Taiwanese and South Korean policy makers to discipline their private sectors and their bureaucracies—an ability to which the label of “strong” or “hard” state is often attached. But where this ability came from, and whether it can be replicated in other settings, remains a mystery. Authoritarianism may have had something to do with the East Asian governance style, but there are too many mismanaged dictatorships around the world to take the hypothesis seriously. One need only look at Sub-Saharan Africa. Krueger is clearly thinking of the East Asian cases when she writes:
The adoption of the same economic policies in response to the same (economic) circumstances will . . . have different consequences under a politically strong leadership of a government with a well functioning bureaucracy capable of carrying out the wishes of the leadership than it will when a weak leadership of a coalition attempts to do the same things in circumstances where bureaucrats believe that they can generate support for opposition to those policies . . . (Krueger 1993, p. 9)
20
Journal of Economic Literature, Vol. XXXIV (March 1996) exceptionally equal by cross-country standards (Table 5). Equality may have been conducive to better governance for at least three different reasons. First, these governments did not generally have to contend with powerful industrial or landed interest groups: therefore, policy making and implementation could be insulated from pressure-group politics. Second, the absence of largescale inequities meant that governments felt no immediate need to undertake redistributive policies; they could concentrate on expanding the pie instead. Note that this point is analytically distinct from the previous one. Even a government which is completely insulated from lobbying groups would wish to redistribute income as long as its objective function looks anything like a conventional social welfare function. This redistribu-
But we do not get much more help than this acknowledgement that governments and bureaucracies differ. It is reasonable to suppose that at least part of the explanation has to do with with some of the special initial conditions that the East Asian countries shared prior to their economic take-off. Two such conditions stand out. First, by the late 1950s the East Asian economies had for the most part a much better educated labor force than would have been expected on the basis of their income levels (Table 4). This may have made it easier to establish a competent bureaucracy (as well as enhancing the productivity of interventions aimed at boosting private investment, as argued in Rodrik 1995a). Second, and perhaps more importantly, in all of them the distribution of income and wealth around 1960 was
TABLE 5 DISTRIBUTIONAL INDICATORS FOR EAST ASIAN AND COMPARATOR COUNTRIES, AROUND 1960 Gini Coefficient for Income, c. 1960 0.49 0.33 0.40 0.34 0.42 0.31 0.40 0.41 0.39 0.44 0.53 0.42 0.42 0.64 0.53 0.45 0.56 0.50 Gini Coefficient for Land Ownership, c. 1960 n.a. n.a. 0.47 0.39 0.47 0.46 n.a. 0.46 0.45 0.87 0.85 0.67 0.52 0.69 0.69 0.53 0.59 0.68
Country East Asia Hong Kong Indonesia Japan Korea Malaysia Taiwan Singapore Thailand Unweighted Average Others Argentina Brazil Egypt India Kenya Mexico Philippines Turkey Unweighted Average Source: Rodrik (1994a)
Rodrik: Understanding Economic Policy Reform tive motive figures more heavily the worse the pre-tax distribution of resources among the population.15 Third, and related to these, the fact that the top political leadership was free to focus on economic goals meant that it could supervise the bureaucracy closely and make sure that the bureaucrats assisted rather than hindered private entrepreneurship. For bureaucracies are prone to two problems that are fatal to economic performance: they can be captured by the interests they are supposed to regulate, and they can create excessive red tape discouraging economic activity. In Korea and Taiwan, these problems were avoided because the bureaucracies were very closely supervised by the top political leadership. These points have recently received some indirect empirical support in a number of papers which have found a positive relationship between equality and subsequent economic growth. See in particular Alesina and Rodrik (1994), Persson and Tabellini (1994), George Clarke (1993), and Nancy Birdsall, David Ross, and Richard Sabot (1994). In these papers, the initial level of income equality around 1960 is shown to be robustly and positively correlated with growth over the next three decades, controlling for other initial conditions such as per capita income and educational attainment. The theoretical models proposed by Alesina and Rodrik (1994) and Persson and Tabellini (1994) to explain this phenomenon rely on a political-economy argument. When distributive policy is sensitive to the preferences of the median voter, it can be shown that the equilibrium level of redistribution is increasing in the gap between the median voter’s income and average income.
15 The analytics backing up this statement can be found in Alberto Alesina and Rodrik (1994) and Torsten Persson and Guido Tabellini (1994). The result is explained later in the text.
21
More equality usually (but not necessarily) goes with a lower difference between median and average incomes. Consequently, and assuming that redistributive policies act as a tax on accumulation, societies with lower inequality will resort to less redistribution and grow faster. The empirical literature to date has not been successful in testing directly for the presumed political channel through which equality affects growth. Therefore the precise links between equality and growth remain to be demonstrated (Roberto Perotti 1992b). As the East Asian cases suggest, it is at least plausible that governance plays some role in this relationship. Nonetheless, this too is an area in need of more research. We need to understand better why broadly similar policies can produce quite different outcomes in different societies. D. The Political Economy of Macroeconomic Cycles The continued existence of some of the policies that make up the “import substitution” syndrome, despite mounting evidence of their inefficiencies, can be understood in distributional terms: such policies redistribute income or rents to favored groups in society. Their beneficiaries—business and labor groups protected from foreign and domestic competition—are naturally resistant to their reform. But what is distinctive about large-scale deficit spending and overvalued currencies is that these policies are by their very nature temporary, and the longer they are pursued the more drastic their eventual reversal must be. Groups that benefit during the upswing of the joyride have to suffer losses during the downswing. Labor, for example, normally gains from expansionary “populist” policies, but also ends up as the greatest loser when the eventual crisis takes its toll on real wages and employment. Businessmen benefit from
22
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 6 CONSEQUENCES OF POPULISM IN PERU Average 1980–84
1985 5.8 0.3 2.4 2.3 158.3 64
1986 9.1 6.0 9.5 13.3 62.9 73
1987 12.9 7.2 7.8 8.3 114.5 79
1988 15.6 7.4 −8.8 −11.5 1722.3 60
1989 9.8 1.0 −10.4 −7.5 2775.3 29
Public sector deficit (% of GDP) Current account deficit (% of GDP) Real GDP growth Real consumption growth Inflation Real wages (1979=100) Source: Lago (1991, Table 9.4).
7.8 3.9 −1.0 −0.4 87.0 95
cheap imported inputs while the currency overvaluation lasts, but they are condemned to take a hit when the inevitable devaluation takes place (or when a large black-market premium makes its appearance). Therefore, short of attributing myopia or irrationality to the main political actors, it is hard to understand why such policies find support in the first place. 16 A fascinating case in point is the experience of Peru under President García (1985–90). Facing a stagnant economy, the new García administration launched a “heterodox program” in August 1985. The main thrust of the program was to boost consumption demand by increasing real wages, subsidizing consumption, and creating public-works programs. The government also instituted a freeze on
16 Political scientists too often ignore this issue, by treating all policies as benefiting some groups necessarily at the expense of others. Therefore the presence or absence of reform is viewed in purely partisan terms. This view is taken to its logical extreme by Bates in his comment on Williamson’s (1994) introduction. He explains reform thus: “Economic technocrats become powerful, and thus reform becomes politically sustainable, when they serve the interests of powerful groups: industries, sectors, or regions of the economy” (in Williamson 1994, p. 32). The fact is that reform— particularly in the macroeconomic area—often benefits the politically influential groups that block it.
prices, interest rates, and the exchange rate. And consumption did boom for a couple of years, raising output alongside it (Table 6). According to Lago, “[p]rior to 1987, private-sector confidence in and support of the government’s economic policy could only be described as being unanimous” (1991, p. 281). However, the process was clearly unsustainable: the public-sector and current-account deficits both rose substantially and foreign reserves were depleted. By late 1988, the economy had collapsed and prices were near hyperinflation levels. Real wages, which had increased until 1988, took a sharp nose dive and in 1989 stood at a third of their 1987 level (Table 6). (For the full account, see Lago 1991.) Peru’s case may be extreme, but it is by no means unique. Observers of the developing world have long been fascinated by the prevalence of such boomand-bust cycles. Ranis and Mahmood (1992) document how governments embark on unsustainable spending booms during periods when resources are temporarily plentiful (due either to improvements in the terms of trade or the availability of external finance), and then resist retrenchment when the inevitable downturn arrives. What is the explanation for these cycles? The answers pro-
Rodrik: Understanding Economic Policy Reform vided in the case-study literature, as in Ranis and Mahmood, tend to be mechanical. Ranis and Mahmood (1992, p. vi) summarize their conception of political economy thus: we accept the concept that the typical LDC state is an instrument of the most powerful— if often myopic—interest groups, the new industrialist class, public or private.
23
The operative term here, for our purposes, is “myopic.” Ranis and Mahmood do not offer an explanation for these policy cycles, save for suggesting shortsighted behavior. At least these authors are explicit about their assumption of myopia. More commonly the attribution of myopia to policy makers and interest groups is implicit. For example, Krueger (1993, p. 19) writes:
Regardless of the form of government—benevolent guardian, predatory authoritarian or bureaucratic, or factional—the political process typically demanded more resources than were available from tax revenues in the early stages of growth.
She then lays out in great detail the corrosive and cumulative long-run consequences of deficit spending and fixed exchange rates. But she considers these consequences as largely unanticipated by the relevant groups. But myopic behavior can take us only so far in understanding these policy cycles. For one thing, the underlying economics is straightforward: it does not require a Ph.D. in economics to realize that spending money one does not have can result in unpleasant consequences.17 For another, people ought to learn from their mistakes. To take one—not altogether extreme—example, since the 1950s Turkey has gone through four full boom-and-bust cycles, one per decade.
17 Of course, that does not prevent some people from trying it anyhow. But we’d like to think of this as the exception rather than the rule.
Because economists hate to give up on rationality (at least until they become older—wiser?—and distinguished), it is not surprising that a small analytical literature has developed around this set of issues. Several recent papers have proposed formal models that generate political outcomes that are inefficient from the standpoint of the politically powerful groups themselves, even though these groups behave rationally and non-myopically. The trick is usually performed by positing some kind of coordination problem among the contending actors. Some of these models focus on providing an explanation for the adoption of unsustainable policies, while some focus on the timing of reform (usually fiscal stabilization). A few attempt to do both. In the near-classic of this field, Alesina and Allan Drazen (1991) show how a stabilization can be delayed, at great cost to the parties involved, thanks to a “war of attrition” between two groups, each of which is uncertain about the costs being incurred by the other group. It is individually rational to wait in this model because the group that caves in first is assumed to bear a larger part of the post-stabilization tax burden. Each group has the incentive to wait and see if the other group will throw in the towel first. Stabilization takes place only when one of the groups figures that it stands to gain more from assuming the cost of stabilization than from waiting another instant to see if its rival will do so instead. In other papers, Alesina and Tabellini (1989), Sule Özler and Tabellini (1991), and Alex Cukierman, Edwards, and Tabellini (1992) have shown how macroeconomic problems can arise from political instability (i.e., expected turnover in governments) as well as polarization. In these papers, governmental “myopia” arises not from a war of attrition, but from the realization by groups in power that they may be replaced by future gov-
24
Journal of Economic Literature, Vol. XXXIV (March 1996) oped by Velasco (1994), who considers a model in which two organized groups treat the state’s resources as a common pool, and decide every period how much to extract from the pool. For simplicity and tractability, the choice is limited to one of two options: One option is to extract a “large” share, which leads to a build-up of government debt and a reduction over time in the state’s net wealth. The other is to extract a “small” share, which leaves net government wealth unchanged. Velasco demonstrates the existence of a switching equilibrium in which the two groups demand large transfers until the public debt stock reaches a critical amount, after which they both switch to the more moderate strategy. The latter is a cooperative outcome maintained by trigger strategies, and stands for the stabilization stage. The intuition is that once government net wealth becomes very low (i.e., once the state’s resources are sufficiently plundered), the benefits of the defection strategy become low compared to its costs (which are modeled here as being constant and independent of net wealth). Finally, Perotti (1992a) builds a model with three groups—the poor, the middle class, and the capitalists—to show how unsustainable policies can emerge as a consequence of a “populist” alliance between the poor and the capitalists. Three key features lead to this result. First, it is assumed that capitalists can insulate themselves from the costs of stabilization in the long run because capital is mobile in the long run (and only in the long run). Second, the poor are too poor to be taxed, so they do not bear any of the costs of stabilization. Third, the poor derive a positive externality from the profits made by capitalists at home. Now suppose that the economy is hit with a negative shock which makes longrun capital flight the dominant strategy
ernments with different ideological or redistributive preferences. More recent papers by Labán and Federico Sturzenegger (1994a and 1994b) generate results with the same flavor, by relying not on asymmetric information or uncertainty but on the dynamics of the flight from the domestic currency once inflation sets in. These papers assume two contending social groups—the rich and the poor—which prefer, at low levels of inflation, not to submit themselves to a costly stabilization. The key dynamic is generated by the presence of a “financial adaptation” technology (i.e., bank accounts in Miami), which is available only to the rich (at a fixed cost), and which leads to both increased inflation and a greater burden on the poor over time. The main result is that at some point the poor may therefore become willing to accept conditions they would have earlier rejected. Guillermo Mondino, Sturzenegger, and Tommasi (1992) generalize this framework to generate the possibility of recurring cycles of inflation and stabilization. In this model, two political groups interact strategically and demand transfers from the government, with the transfers paid for by the inflation tax. In addition, members of each of these groups have access to financial adaptation, and they behave atomistically in deciding when to indulge in it. Because the demand for transfers is determined at the group level while the flight from domestic currency is decided at the individual level, socially suboptimal outcomes and cycles both become a possibility. Under low inflation it may be optimal for groups to demand high levels of transfers, which sets into motion rising levels of inflation through the flight from local currency. In turn, at high levels of inflation, it may be optimal for the groups to agree on a stabilization. A variant along the same lines is devel-
Rodrik: Understanding Economic Policy Reform for capitalists. Because domestic capital will be abroad in the long run anyhow and the burden of stabilization does not fall on the poor, the latter’s well-being is maximized by expansionist policies that increase domestic profits in the short run. Hence the poor and the capitalists ally themselves against the middle class in pursuit of unsustainable policies that they know will be reversed eventually. In these models, political agents or groups are assumed to be rational and forward-looking, with expectations that are consistent with the properties of the underlying model. Behavioral rules are derived from solving optimization problems with well-defined objective functions. The political-economic outcome is derived as a Nash equilibrium in which each individual or group is doing the best it can given the actions of others. Of course, none of these models is capable of providing the full political-economic story of the boom-and-bust cycles discussed above. But they each capture an aspect of the phenomenon, and perhaps more importantly, they confirm that we can do better than resort to myopia or irrationality when explaining social phenomena. There are several directions in which this literature could be usefully extended. First, we now need theoretically informed case studies—or more formal test—that attempt to discriminate among these alternative stories. Second, the normative implications of these models for policy and institutional design have to be worked out. Third, there are several relevant aspects of reality which have yet to receive attention. For example, in many political organizations (parties, unions, etc.) principal-agent problems prevent the leadership from internalizing in full the interests of the rank-and-file. It is plausible that this would be one source of inefficiency in
25
economic policy making.18 Similarly, there is no room in these models for learning about the way the economy functions: if it is people that learn rather than organizations, each generation will repeat the mistakes of the previous ones.19 Finally, and most fundamentally, these papers leave hanging a key question: if distributional struggles are at the heart of inefficient policy choices and macroeconomic policy cycles, why do policy makers not design compensation schemes to neutralize political opposition? In the Alesina-Drazen (1991) model, for example, the distribution of the costs of stabilization is taken as exogenous. This is unsatisfactory because the design of the stabilization package can surely influence the distributional impacts. I will return to the issue of compensation toward the end of the paper. III. Crisis and Reform During the 1980s The 1980s experienced two events of lasting significance. First, much of the developing world became engulfed in a protracted debt crisis. Second, many countries began to shed their importsubstitution policies and endorsed market-oriented ones. The reforms did not come immediately after the crisis (Table 7). In fact, the typical pattern was for governments to respond to crisis by tightening their restrictions. Even Chile, which had already opened up during the 1970s under General Pinochet, initially chose to increase its (uniform) import tariff when the crisis hit. But after some delay, which differed across countries, countries in Latin America, Africa, and Asia jumped on the bandwagon of re18 John Pencavel has suggested that the policies of the Mexican petroleum labor union at PEMEX can be explained in such a framework. 19 The relevance of these considerations was suggested by Pencavel in his comments on this paper.
26
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 7 TIMING OF REFORMS AND INFLATION Annualized inflation rate prior to stabilization (%) 23,455 159 1,344 12,378 2,103 16
Stabilization Boliva Mexico Argentina Peru Brazil India 1985 1987–88 1991 1990 1994 1991
Structural Reform 1985 1985–1988 1987–1991 1990 1988–1990 1991
form. Governments endorsed reform with varying degrees of enthusisasm. The most enthusisastic reformers by far were in Latin America, where the Chilean example played a major role. (For more on Latin America, see Edwards 1994.) The developing countries were followed after 1989 by the socialist countries of Eastern Europe and the former Soviet Union, which embarked on their own transition to market economies. A. Does Crisis “Cause” Reform? It is natural to suppose that crisis and reform were somehow related. Indeed, if there is one single theme that runs through the length of the political economy literature it is the idea that crisis is the instigator of reform. This theme keeps reappearing in different guises. Ranis and Mahmood write that resistance [to] vested interests can be overcome only when the system has no other way of avoiding the required adjustment. In other words, it has to be “up against it,” without any easy alternative by which the day of reckoning can be delayed—either because a country’s own resources or the resources made available from the outside are considered insufficient to keep the system operating along the old rails. (1992, p. vi)
“When populist leaders in Argentina, Bolivia, Venezuela, Peru, and Brazil adopted nonpopulist policies,” claims
Bresser Pereira, “it was because the crisis in these countries was so deep that even the costs of sticking to populist policies became higher than the costs of adjustment” (1993, p. 57). According to Krueger, economic reforms were undertaken when “economic conditions deteriorated sufficiently so that there emerged a political imperative for better economic performance” (1993, p. 109). Moreover, the idea is not limited to developing countries. “The ‘golden rule’ of recent Spanish economic history,” writes Guillermo de la Dehesa “is: ‘Only when the level of reserves was sufficiently low and/or the current account balance was in large deficit have necessary economic adjustment and structural reform measures been taken’ . . .” (in Williamson 1994, p. 137; footnote omitted). Explaining why Australia undertook fewer reforms than New Zealand, Max Corden writes “[t]he reforms have been less dramatic than New Zealand’s because things never got so bad: inflation did not rise so high, and considerable trade liberalization had taken place since the 1950s” (in Williamson 1994, p. 112). The idea that governments have to be up against the wall before they act may be valid, but it is not entirely free of problems from an analytical standpoint. First, note that there is a strong element of tautology in the association of reform
Rodrik: Understanding Economic Policy Reform with crisis. Reform naturally becomes an issue only when current policies are perceived to be not working. A crisis is just an extreme instance of policy failure. That policy reform should follow crisis, then, is no more surprising than smoke following fire. 20 Furthermore, the hypothesis is virtually nonfalsifiable: if an economy in crisis has not yet reformed, the frequently proffered explanation is that the crisis has not yet become “severe enough.” Bresser Pereira, who was briefly Brazilian finance minister in 1987, argues that the reason his fiscal package was not supported by President Sarney was that the crisis was not yet perceived as serious (in Williamson 1994, pp. 333–54). True, perhaps, but not very helpful for explanation or prediction. What we surely need to understand is why South Korea’s politicians are ready to change course at the slightest hint of a crisis, while Brazil’s will bring their economy to the brink of hyperinflation several times before they tackle the problem. The political-economy literature recognizes this issue, but is largely silent on it. Some helpful hints come from the analytical literature mentioned previously. Alesina and Drazen (1991) show in their “war of attrition” model that the lower the “degree of cohesion” in society, which they model as the expected asymmetry in the burden of stabilization, the greater the delay before a stabilization takes place.21 One can loosely interpret this in the following terms. In societies where resources are evenly distributed, and the government has a good distributional track record, groups are less in20 Williamson and Haggard (in Williamson 1994, p. 564) mention several cases where reform was undertaken despite the absence of crisis (notably Australia, Colombia, and Portugal). 21 This is also possibly what Krueger has in mind when she argues that a “strong” government such as Korea’s will react quickly to crisis, while a “factional” one will delay (Krueger 1993, p. 126).
27
clined to believe that the burden of adjustment will be one-sided. This places the East Asian countries in an advantageous situation, once again, where crises are concerned. In addition, the emphasis on crisis has in itself little predictive content as to what form the response will take. Latin American countries eventually adjusted not only by balancing their fiscal accounts, but also by overhauling their trade and industrial policies and undertaking privatization on a major scale. As I have discussed above, these trade and industrial policies had little to do with instigating the crisis, and consequently their reform was not a logical necessity once decisive action was taken. In fact, some of the reforms in the area of trade liberalization almost surely complicated the macroeconomic stabilization effort.22 Therefore, the comprehensiveness of the reforms that the Latin American countries undertook still requires explanation. Part of the explanation has to do with the advice that these governments received. As mentioned previously, many professional economists themselves tended to lump together all “import substitution” policies and to hold them equally responsible for the crisis. That was certainly the approach taken by the World Bank, the most important conduit of economic ideas to developing-country policy makers. But how could these wide ranging
22 That is so for several reasons. The reduction of tariffs resulted in losses in fiscal revenue. Furthermore, trade liberalization aggravated the adverse effect of stabilization policies on aggregate demand and domestic output. Last but not least, it created an important conflict in exchange-rate management: trade liberalization required a “compensating” devaluation of the currency, which often could not be undertaken for fear of upsetting fragile price expectations. The recent Mexican debacle could possibly have been avoided, or at least its impact minimized, if the authorities had not removed controls on trade and capital flows so enthusiastically.
28
Journal of Economic Literature, Vol. XXXIV (March 1996) ciated with this reform can be expressed as 1/µεt, where µ is the share of affected imports in domestic consumption and ε is the absolute value of the import demand elasticity. With reasonable values of t = 0.5, µ = 0.2, and ε = 2, the formula yields a value of five for the PCBR. 23 That is, five pesos of income have to be reshuffled among groups in society to yield a single peso of efficiency gain! This is one way of capturing the political difficulty of achieving microeconomic reform in normal times. Now suppose that the economy gets caught up in a deep macroeconomic crisis, recovery from which would provide all-around gains. The idea here is that triple-digit inflation and the associated instability hurt pretty much all groups in society, although some may naturally protect themselves better than others. Suppose further that a reformist government packages the stabilization program needed to end the crisis with the price reform discussed above. Let γ stand for the percentage increase in income that accrues to all groups in society as a result of the stabilization. We can think of γ as an index of the depth of the macroeconomic crisis. Under reasonable assumptions on the value of γ (e.g., γ = 0.10) the PCBR plummets to well below unity (from five previously). The point is that the opportunity to do something that will benefit most everyone by a large margin—an opportunity that arises only when the economy is mismanaged terribly and falls into deep crisis—allows reformist policy makers to sneak in, along23 There are three groups that matter in this example, consumers of the importable, producers of the importable and the rest of the economy as affected by the government budget. Let the world price of the importable be p*, the domestic price p, and the quantities of consumption, production, and imports c, q, and m, respectively. Consumers gain -cdp, producers gain qdp, and the government budget changes by mdp + tp*dm. The efficiency gain is tp*dm. See Rodrik (1994b).
trade and industrial policy reforms be rendered palatable to the interest groups that had been their beneficiary for so long? How were they persuaded to go along? Here again it is plausible that the atmosphere of crisis played a role. Crisis enabled reformist governments to package fiscal reforms, which were absolutely crucial for the return of price stability, with trade and industrial policy reforms, which were viewed as desirable in the longer run but were incidental to the immediate crisis. In other words, policy makers acted as agenda setters: They presented domestic interests with a package of both macroeconomic and microeconomic reforms. Because high inflation and macroeconomic instability hurt pretty much everyone across the board, influential interest groups felt compelled to go along. They may have preferred to have only the macroeconomic component of the package, but that was not the choice that they confronted. The argument can be illustrated by using a simple heuristic concept that we may call the “political cost-benefit ratio” (PCBR, Rodrik 1994b). Assume that redistributing income is politically costly, and that this cost has to be traded off against the efficiency benefits of reform. Define the PCBR as the ratio of the total redistribution generated by reform to its efficiency benefits. This index answers the question: how many pesos are reshuffled among different groups for every peso of efficiency benefit? The PCBR takes values between zero and infinity. The higher the value of the index, the more difficult reform is likely to prove. Price reforms (such as trade liberalization and removal of subsidies) tend to generate a lot of redistribution relative to their efficiency benefits. Consider the removal of an import tariff, t. To a firstorder of approximation, the PCBR asso-
Rodrik: Understanding Economic Policy Reform side the stabilization, microeconomic, and structural reforms which have significant distributional implications and which would be difficult to implement under normal circumstances. B. Does Reform Have Short-Term Costs? The proposition that some of the major structural reforms of the 1980s and 1990s were adopted despite interest groups, rather than because of them, naturally raises the question of the sustainability of the reforms. Here we face two other conventional wisdoms of the political-economy literature: First, reforms become sustainable when they generate “winners” with a stake in their continuation. Second, and this one is the downer, reforms tend to make things worse before they make them better. For a proposition that is startlingly lacking in empirical support, the second piece of conventional wisdom is surprisingly strongly held. It permeates practically every discussion of the political economy of reform. “The results of many worthwhile reforms lie on a J-curve,” writes Piñera, “they tend to make things a good deal worse before they get better” (1994, p. 227). Bresser Pereira, Maravall, and Przeworski (1993, p. 2) state categorically: “the effect of economic reform on growth must be negative in the short run.” “A key political problem of sustaining support for reform programs,” argues Joan Nelson “is the long delay in reaping visible benefits for much of the population.” She continues: “Quite simply, in most cases there are not enough early winners to ensure the political sustainability of the program . . .” (in Williamson 1994, p. 476–77). The facts do not support such pessimism about long-delayed response. Once one makes allowance for the likelihood that the counterfactual—no reform— produces even worse results in the short
29
run, the consequences of reform actually look pretty good. This shows up in a number of different areas. With regard to disinflation, most of the recent cases of exchange-rate based stabilizations, such as those in Israel (1985), Mexico (1987), and Argentina (1991), have been accompanied by consumption booms, rather than recessions.24 With regard to broader structural adjustment policies, the best statistical evidence to date is that such policies tend to significantly increase, rather than decrease, growth of output within two or three years, even if not immediately (Corbo and Patricio Rojas 1992).25 Even in Eastern Europe, where there is genuine reason to believe that the transition may have short-term costs, the evidence indicates that reform reduces rather then intensifies the shortterm costs. Output has fallen the least in countries like Poland and the Czech Republic which have had the most extensive reforms, and the most in countries like the Ukraine which have had the least (Table 8). Furthermore, consumption statistics indicate that much of the measured contraction of GDP in transitional economies is illusory. (For the Polish case, see Andrew Berg 1993.) Of course, structural reforms (and possibly stabilization as well) may have sharp distributional consequences, as discussed previously. But this is not a matter of the short run versus the long run, as with the worries reflected in the quotes above. Saying that reform is diffi24 See for example Guillermo Calvo and Carlos Végh (1994). Dornbusch and Fischer (1993) argue that it may be unavoidable to incur output costs when disinflating from moderate inflation (40–80 percent per year), but their evidence on this is not so clear cut. 25 Corbo and Rojas use a control-group approach and focus on the 1985–88 performance of countries that received at least two Structural Adjustment Loans (SAL) from the World Bank starting in 1985 or before. They find that adjustment raised GDP growth rates by an average of two percentage points.
30
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 8 REFORM IN ECONOMIES IN TRANSITION Average annual real GDP change during first three years of transition (%)* Countries with Strong Reform Programs Poland −5.5 Czech Republic** −8.2 Hungary −6.2 Estonia −10.1 Countries with Weak Reform Programs Russia −14.7 Ukraine −14.4 Moldova −17.8
Annual inflation, 1994(%) 30 9 19 47 336 1000 245
Source: Calculated from IMF (Oct. 1994, Table 11). ∗ The first year of the reform program is taken to be 1990 for Poland and Hungary, and 1991 for others. ∗∗ The figures for Czechoslovakia are used for 1991 and 1992.
cult because some powerful groups will be inevitably made worse off—in the long run as well as the short run—is distinct from claiming that reform is difficult because the net benefits from the reform come too late for politicians to reap the gains. It is the latter that seems empirically problematic. None of this is to dismiss the obvious fact that reforms do arouse opposition, and that the opposition often tends to be strongest early on. The point is that we cannot really attribute these phenomena to presumed (rather than demonstrated) short-term costs of reform. Once irrationality and myopia are dismissed, the political economy of reform turns out once again to contain more puzzles waiting to be worked out. C. Does Foreign Aid Help Reform? If reform has short-term costs, as so many believe, then foreign aid should help reforms get launched (and sustained) by alleviating these costs. This indeed is the standard justification for the World Bank’s and the IMF’s “struc-
tural adjustment” lending as well as for official bilateral credits. But the logic of this statement is not unassailable, even if the premise is accepted. The reason is that external resources reduce the costs both of reform and of doing nothing— that is, avoiding reform. In addition, the prospect of aid can actually exacerbate the delay in stabilization, by inducing groups to postpone making sacrifices until aid actually materializes. 26 The effect on reform is consequently ambiguous. One of the strongest proponents of foreign aid has been Jeffrey Sachs, who has served as a high-profile advisor to reformist governments in Bolivia, Poland, and Russia among other places. In his contribution to Williamson (1994) entitled “Life in the Economic Emergency Room,” he complains that economists tend to neglect the role played by foreign assistance in most of the major postwar reforms. He provides capsule histo26 This argument is worked out in Alessandra Casella and Barry Eichengreen (1994), using the Alesina-Drazen model.
Rodrik: Understanding Economic Policy Reform ries of the economic reforms in postwar Germany, Japan, Bolivia, Mexico, Chile, Poland, Israel, and Turkey, stressing the critical contribution of aid in each case. Sachs is explicit that a government committed to reform is needed before aid can do any good. But once such a government is in place, the international community should be forthcoming. What aid can do, even if it is small in relation to the country’s problems “is help good governments to survive long enough to solve problems” (in Williamson 1994, p. 512). The problem is that aid can also help bad governments survive. For debating purposes, one can cite at least as many cases as Sachs does to demonstrate an association between plentiful aid and delayed reform. 27 As mentioned previously, one of the themes of the book by Ranis and Mahmood (1992) is that the availability of external resources tends to promote irresponsible policies. Scarcity of resources, on the other hand, is good for reform. (Note the obvious parallel to the crisis hypothesis.) One of the pieces of conventional wisdom about the Korean and Taiwanese reforms of the 1960s is that these reforms took place in large measure because U.S. aid, which had been plentiful during the 1950s, was coming to an end (see for example Ranis and Mahmood 1992, p. 139). Sachs would respond that donors must ensure the recipient governments will undertake the reforms before doling out the cash. More generally, aid must come with a heavy dose of conditionality. This is well recognized, and both the IMF and the World Bank make access to their resources conditional on good behavior
27 Some of Sachs’ case histories can also be challenged. It is true that aid played an important role in the Turkish reforms of the early 1980s, but it is also true that it delayed the fiscal day of reckoning, with ultimately debilitating consequences for the Turkish economy.
31
on the part of the borrowing governments. But conditionality is no panacea, as conditionality can last (at best) only as long as net transfers are positive. In the end, sovereign governments are just that: sovereign. There is a fairly large literature on conditionality and its effectiveness (see Manuel Guitian 1982; Paul Mosley 1987; Jacques Polak 1991; and V. Thomas et al. 1991). Some emerging conclusions are that conditionality is most effective when it focuses on a small range of quantifiable indicators and when governments are already committed to the reforms. However, it may not be easy to tell governments apart, and even governments that are genuinely reformist may be tempted by aid to deliver less than they promised. IV. In Search of a Manual for Reformist Politicians All of which raise the question, “what’s a poor reformist politican to do?” Recently Williamson had the interesting idea of gathering a group of high-ranking technocrats (“technopols” in Williamson’s jargon) to talk about their experiences, in the hope that some common lessons may emerge. The resulting book (Williamson 1994) opens with a list of hypotheses drawn from the literature about what makes reform feasible and successful, and which contributors were asked to examine from their own individual perspective. I have reproduced the list in Table 9. The list includes some items that I have already discussed, as well as many others. In the end, however, Williamson is forced to concede defeat. In his concluding chapter (written with Haggard), he writes “there are no fully robust empirical generalizations; in every case there is at least one partial counterexample. None of the 15 hypotheses investigated was either necessary or sufficient for
32
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 9 HYPOTHESES ABOUT REFORM
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
Policy reforms emerge in response to crisis Strong external support (aid) is an important condition for successful reform Authoritarian regimes are best at carrying out reform Policy reform is a right-wing-program Reformers enjoy a “honeymoon period” of support before opposition builds up Reforms are difficult to sustain unless the government has a solid base of legislative support A government may compensate for the lack of a strong base of support if the opposition is weak and fragmented Social consensus is a powerful factor impelling reform Visionary leadership is important A coherent and united economic team is important Successful reform requires economists in positions of political responsibility Successful reform requres a comprehensive program capable of rapid implementation Reformers should mask their intentions to the general public Reformers should make good use of the media Reform becomes easier if the losers are compensated Sustainability can be enhanced by accelerating the emergence of winners
Source: Williamson (1994)
successful reform” (p. 589). 28 The propositions that received strong support were “the need for a strong political base, for visionary leadership, and for a coherent economic team.” Disproved were the hypotheses that reform requires authoritarian regimes or that it is an exclusively right wing enterprise. But the volume itself is far from a failure. The exchanges among former politicians, technocrats, economists, and political scientists that it contains make for fascinating reading. A fundamental fault line that divides the contributors to this literature is the issue of how participatory reform politics ought to be. Most economists are on the side of speed, stealth, and consequently of reform from above. Vladimir Mau, a Russian economist, writes:
28 As indicated in Table 9, there were actually 16 hypotheses, so the number 15 must be a misprint.
Count Sergei Witte, a prominent reformer under the czars, used to say that there were two essential elements for radical reforms in Russia: absolute monarchy, because you need not pay attention to your critics if His Majesty supported you, and speed, because somebody might persuade the czar to change his mind before the reform could be made irreversible. (pp. 435–36)
Sachs cites approvingly the example of Jacques Necker (the French finance minister before the 1789 revolution) who maintained that flexibility or willingnesss to compromise, which might be harmless or even advantageous in other ministers, was an unforgivable failing in a finance minister!
Sachs goes on:
If reformers want free prices, they should not stand around and talk about it—they should do it, because everyone will be against freeing prices until it has been done, until it is an established fact. (pp. 509–10)
Sachs has plenty of first-hand experience in Bolivia, Poland, Russia, and elsewhere, so his views perhaps ought to
Rodrik: Understanding Economic Policy Reform count for a lot more than those of armchair economists. He is especially outspoken on the question of whether reformers should strive for a consensus with affected groups.29 “Many participants have suggested that reformers succeed by constructing a ‘social consensus’ in favor of reforms,” he writes. “This is mostly not the case. In deep crises, there is simply no consensus to build upon, only confusion, anxiety, and a cacophony of conflicting opinions.” He goes on to relate how the Polish “big bang” was enormously controversial at the time, and the desire to “return to Europe” did not translate into agreement on specific policies. “Consensus on many specifics (e.g., currency convertibility, price decontrol, budgetary discipline) has come to Poland, but only after three years of reform” (p. 505). In Sachs’ view, it is at best a waste of time to seek a broad coalition for reform because most people have no understanding of what is required: “While the history of marketbased reforms has repeatedly shown that free markets, open trade, and an economy fueled by private ownership are enormously powerful in stimulating rapid economic growth, the general public rarely knows it or believes it at the start” (p. 506). Neither is the problem limited to the person on the street: “Few Russians understand the source of Russia’s current inflation, least of all the [former] governor of the central bank . . .” (p. 507). The operational implication of all this is that reform needs a strong and autonomous executive, unhindered by the search for consensus and compromise. 30
29 Of course, there are very few things that Sachs is not outspoken about. His sharp critique of IMF policy in Russia in his after-dinner speech at the Williamson conference drew a defensive response from the IMF (printed in the Williamson volume), and (naturally) a rejoinder from Sachs. 30 In a personal communication, Sachs has clarified that his views should not be taken to imply
33
Sachs’ words have the advantage of expressing forthrightly what a lot of economists feel deep down but find politically incorrect to articulate. What some may find striking in these statements is the lack of faith in the common sense of ordinary people and in the efficacy of political institutions, especially in new democracies. That economists should hold to these views is not without irony. After all, homo economicus is supposed to be rational and forward-looking, and to process all the information that comes his way in the most efficient way. Sachs’ homo politicus is none of these things, or else is being held hostage to some grand coordination failure whose nature is unclear.31 This irony is seized upon by Przeworski, who provides a serious challenge to the economists’ preference for reform from above (in Bresser Pereira, Maravall, and Przeworski 1993). Focusing on the Polish case, he faults the storm tactics favored by Sachs and others for both weakening democratic political institutions and making errors in economic policy more likely. He and his two coauthors (Bresser Pereira and Maravall) summarize their argument in their concluding chapter thus: we find that subjecting the reform strategy to the competitive interplay of political forces is superior on three essential grounds: It
that he favors authoritarian forms of government over democracy. He believes, however, that democratically elected governments should have a fairly free hand in between elections. 31 One reader objected to this paragraph in the following terms: “The notion that the conventional economist may be a better judge of economic policy than the man on the street contains no more ‘irony’ than the notion that a physician may know better how to diagnose and cure my illness.” Most of us, however, take our physician’s advice. Were economists’ advice as well received by politicians, the puzzles on which this paper focuses would be far less salient.
34
Journal of Economic Literature, Vol. XXXIV (March 1996) a terrifying prospect for most Poles. 32 The main shortcoming of the Balcerowicz plan, therefore, was its failure to put in place adequate social protection to guard against the inevitable unemployment that would result. 33 Przeworski argues that this was a “technical” mistake, but one that resulted from not having listened to the people:
If the purpose of their architects was to make reforms politically palatable, the blueprint was not well designed. But the reason technocrats commit ‘technical’ errors is that they do not consult and concert with those who are affected by their blueprints. There is something paradoxical when believers in the informational efficiency of decentralized decisions fear them most . . . The main obstacle to reform is people. (p. 185)
32 That the decline in popularity was related to rising unemployment is plausible enough, but the statistical analysis that Przeworski undertakes to establish this relationship is unconvincing. The analysis consists of pairwise correlations between poll results and actual economic outcomes. The results of the exercise are unreliable for several reasons: there are few observations (a maximum of 20), there is obvious scope for omitted-variable bias, and a lot of the variables have strong time trends leading to spurious correlations. The strong negative correlation Przeworski obtains between the program’s popularity and unemployment is a statistical reflection of the fact that the first of these consistently went down, and the second up, after January 1990. A more telling piece of evidence comes from the experience of the Czech Republic (not discussed by Przeworski). Alone among reformist governments, Vaclav Klaus’ government has retained much of its popularity. Not coincidentally, the Czech Republic is also the only strongly reforming economy that has not (yet) experienced high unemployment. An alternative argument for the decline in the popularity of reforms is provided in Rodrik (forthcoming), based on changes in the preferences of state-sector workers over the course of the transition. 33 This point would be disputed by Balcerowicz and Sachs. According to Sachs (1994, p. 3): “Contrary to the impression of uncompromising laissez faire and harsh social policies, Poland’s social spending has risen very sharply since the beginning of reforms. Total budgetary spending on social programs (labor fund, pension fund, health care) has risen from 10 percent of GNP in 1990 to 21 percent of GNP in 1993, one of the highest proportions in Europe.”
improves policy, it builds support for the continuation of reforms, and it helps consolidate democratic institutions. We do not see a trade-off between public discussion and the soundness of economic plans. (p. 210)
These points are underscored in Przeworski’s interesting discussion of the Polish program. The program began with the famous “big bang” of January 1st, 1990. Leszek Balcerowicz, a contributor to the Williamson volume (and from whom more later), was its principal architect and Sachs its principal foreign advisor. Przeworski grants that the nearhyperinflation in which the Polish economy found itself by the Fall of 1989 required some fast action. But he faults the program for having had no social dimension: “I could find no references to building a welfare state or extending social protection in any formulation of the reform program,” he declares; “this was a pure trickle-down model of reforms” (p. 142). This may be bad enough for some, but the real clincher in Przeworski’s argument is that the absence of social policy ended up threatening the sustainability of the economic reforms. To demonstrate this, he analyzes at length the trend in public support for the Balcerowicz program. He finds that until May 1990 public support was high and steady, after which a significant decline in support occurred. Thereafter it remained steady again, until early 1991 when “the diagnosis of the current situation, optimism, willingness to bear sacrifices, and support for the Balcerowicz program all began to decline again” (pp. 158–59). Eventually, of course, support would fall so low that the former Communists would return to power in the September 1993 parliamentary elections. Przeworski attributes this slide in support primarily to the rise in unemployment, which he argues convincingly was
Rodrik: Understanding Economic Policy Reform
Moreover, Przeworski and his colleagues believe that the top-down style of these reforms tend to corrode nascent democratic institutions: the autocratic policy style characteristic of Washington-style reforms tends to undermine representative institutions, to personalize politics, and to generate a climate in which politics becomes reduced to fixes, to a search for redemption. Even if neo-liberal reform packages make good economics, they are likely to generate voodoo politics. (pp. 9–10)
35
Przeworski undertakes an interesting analysis of the trend in public confidence in various Polish institutions. He documents how confidence in representative institutions like political parties, the government, and the Sejm gradually eroded while confidence in the army and the police increased, to the point where “two years after the transition to democracy, Poland was a country in which the three institutions in which people had the most confidence were the army, the church, and the police” (p. 173). He attributes this to the “policy style” of the reforms— the introduction and continuation of a set of policies “by surprise and independently of public opinion and of representative organizations and institutions” 34 (p. 175). To paraphrase Przeworski only slightly, economic reform is consequently too important to leave to economists. The alternative to which Przeworski and his colleagues would subscribe is not entirely clear. In terms of actual policy content, their “social democratic” approach departs only mildly from the Washington consensus, in envisioning a somewhat more active role for the government in industrial policy and in coordinating private investment. 35 The rest
34 “The televised spectacle of the government ramming through the parliament a package of complicated legislation . . . took its toll” (p. 176). 35 To which Sachs would retort: “Governments that have reached hyperinflation cannot self-
of the program, including the emphasis on a “social safety net” (to use the Washington term), would not look out of place in a World Bank document. Przeworski mentions the Hungarian case approvingly, and implies that the more gradual approach adopted there was preferable because it paid greater attention to social issues. Yet, Hungarian unemployment has risen nearly as high as the Polish one, and by 1994 most economists would agree that the Polish economy had turned the corner while Hungary was still struggling. The cumulative output fall in Hungary has been larger, and the recovery smaller, than in Poland (see John Flemming 1995, Table 4). In any case, in 1994 former Communists were elected to power in Hungary as well. Perhaps an alternative model would have been the “social pacts” approach, used most conspicuously by post-Franco governments in Spain. This experience is described in Maravall’s contribution to the same volume (as well as by de la Dehesa in the Wiliamson volume). The Moncloa pacts in Spain were an effective device through which labor received expanded rights in return for wage moderation. But as Maravall makes clear, this experience is probably of little relevance to Eastern European countries, because there was still substantial room for expanded government in Southern Europe. Regarding the experience of Spain, Portugal, and Greece he writes:
Both public expenditure as a whole and social expenditure in particular increased very sharply from the mid 1970s under the new democratic regimes. Welfare systems and mechanisms for income distribution were extended and reorganized. The role of the state evidently, be expected to develop complex industrial policies or structural policies. After all, they aren’t even carrying out their most fundamental task” (in Williamson 1994, p. 510).
36
Journal of Economic Literature, Vol. XXXIV (March 1996) accompli than it would have been to build reforms gradually in the face of strong populist opposition in Parliament after the elections of October 1991. (pp. 172–73)
in the provision of health, education, and pensions was reinforced. Revenues from taxation increased, and fiscal systems were made more redistributive. (Bresser Pereira, Maravall, and Przeworski 1993, p. 101)
As he later puts it, “public expenditures could generally grow in spite of crisis” (Bresser Pereira, Maravall, and Przeworski 1993, p. 105). Leszek Balcerowicz, the former deputy prime minister after whom the Polish reform program was named, agrees with Przeworski’s facts, but not with the conclusions. There was no choice but to move quickly on both the macroeconomic and structural fronts, and public debate and deliberation would have wasted valuable time: “[T]he establishment of a social pact [as in Spain] in Poland in 1989 would simply have taken too much time during a crucial period.” Besides,
[w]hat would be the purpose of such a pact today? It could not afford to give compensation, because that would conflict with the fiscal program; the real compensation had consisted of the rapid elimination of shortages, which has been increasingly appreciated as time has passed. (in Williamson 1994, p. 218)
Moreover, it would have allowed the opponents of the program to dilute it. Like Sachs, Balcerowicz has little confidence in the political process in the short term, and complains that two election campaigns in the first two years of the program—presidential in 1990 and parliamentary in 1991—made his job extremely difficult. He sees the first few months of a honeymoon period in early 1990 as having provided an opportunity to ram through as many reforms as quickly as possible. “[T]he period of ‘extraordinary politics’ was short-lived and . . . one should use it to introduce tough economic measures.” Further:
[i]t was easier for the supporters of the market-oriented reforms to defend them as fait
One useful distinction in this debate, not often made, is between the initiation and consolidation of reform. Haggard argues that the initiation of reform requires independence or autonomy for the executive, while consolidation of reform necessitates “building of legislative and interest-group bases of support” (in Williamson 1994, p. 468). This argument is echoed by Joan Nelson, who says that it may be inevitable to be somewhat autocratic in the early stages of reform, but that in later stages one needs to generate support and consensus (pp. 472ff.).36 But what if the autocratic approach early on does irreparable damage to confidence in democratic institutions, as Przeworski claims it does? It is difficult not to feel sympathy for Przeworski’s yearning for a more democratic style of reform. Yet my suspicion is that most observers of the Polish reforms would agree with Balcerowicz and Sachs that speed (and stealth) was of the essence. While there have been some reversals in the reforms, even with these reversals the Polish economy is in far better shape today than it would have
36 Perhaps Mexico (prior to its most recent crisis) is a good example of the Haggard-Nelson strategy. José Córdoba relates in detail the dramatic changes in Mexican economic policy following the 1982 crisis. An astounding turnaround in the fiscal stance (of 10 percentage points of GDP!) was accomplished within two years, and in a fairly unilateral manner. But the government’s subsequent approach was more participatory (or at least, corporatist): “The Economic Solidarity Pact [of 1987] clearly established that inflation could not be reduced through government policies alone but required a social dialogue and active support by labor, business, and peasant organizations. That lesson profoundly impressed policymakers, and from then on much greater effort was made to build consensus and promote the direct involvement of those sectors in society that would be affected positively or negatively by the reforms” (in Williamson 1994, p. 254).
Rodrik: Understanding Economic Policy Reform been under a more gradual path. Balcerowicz is absolutely right that his reforms have created a new, largely irreversible reality that his successors dare not touch—a matter of importance when these successors are former Communists. This fascinating debate takes us back to the most fundamental issue in the political economy of policy reform, the one with which this essay opened: If the objective of reform is to make people better off, why does reform have to be shielded from the people? “[T]he sad fact that good things—such as democracy and market-oriented economic policy—do not always go together,” as Haggard puts it (in Williamson 1994, p. 467), suggests that there is a real puzzle here. Furthermore, if the problem with reform is that powerful groups are hurt by it, why can’t policy makers come up with compensation schemes that remove the hurdle? Myopia, to which many observers ultimately resort, appears to me to be as unsatisfactory an explanation here as in any area of conventional economics. As I pointed out above, the analytical literature has begun to address these puzzles. I close this section by discussing briefly two papers which are especially germane to the questions at hand. The first one (Raquel Fernandez and Rodrik 1991) provides an answer to why reforms that will benefit most everyone may still be rejected by a majority of the electorate, and why compensation may not remove distributional obstacles. The second (Mathias Dewatripont and Gerard Roland 1992) discusses optimal reform design under dynamic political constraints, restricting feasible policies to those that for example will be approved by unanimity. In Fernandez and Rodrik (1991) we ask whether a rational electorate would ever reject a reform which is known to
37
benefit a majority of the voters. We show that the answer is “yes.” We also show that political systems in general have a bias toward the status quo even when the status quo is inefficient and individuals are risk neutral. The key to the argument is uncertainty of a particular kind: the identity of many of the gainers (as well as losers) from reform cannot be determined ex ante. To see how the argument works, consider a democracy where a majority vote is needed before reform can be adopted. Let the economy have 100 voters and suppose that the reform in question will increase the incomes of 51 individuals by five zlotys each and decrease the incomes of the rest by one zloty each, leaving a net gain of (5 × 51) − (1 × 49) = 206 zlotys. In the absence of uncertainty, the majority of the population would vote in favor and the reform would be adopted. Assume that all these consequences of reform are common knowledge. Now suppose that while 49 individuals know for sure that they will gain, the remaining 51 are in the dark as to which among them will gain and which will lose. However, because aggregate consequences are common knowledge, individuals in the latter group know that two of them will eventually benefit while 49 will lose out. (Such uncertainty may arise, say, from incomplete information at the individual level about the skills needed to suceed in the post-reform environment.) This renders individuals in the second group identical ex ante, with an expected benefit from reform of [(5 × 2) − (1 × 49)]/51 = − 0.76 zloty each. Hence the individuals in the uncertain group will reject reform, blocking its adoption. The bottom line is that uncertainty about the consequences of reform at the level of the individual can prevent reform, even when it is recognized that reform will make a politically effective ma-
38
Journal of Economic Literature, Vol. XXXIV (March 1996) tial reform, which generates less of a budgetary burden. Dewatripont and Roland show that a dynamic version of this set-up produces an argument for gradualism. A partial reform early on would induce only the highest-ability group to leave; next period it becomes optimal to induce the next-highest group to leave; and so on, at much lower cost to the budget ultimately. The government’s assumed control over the agenda of policies under consideration allows it to dynamically relax the political constraints. A somewhat similar story, but using the FernandezRodrik (1991) model and based on the Chinese example, is provided by ShangJin Wei (1992), who shows how a sequence of reforms in different policy arenas may be politically sustainable where an economy-wide “big bang” would fail. V. Concluding Comments A political scientist or historian may well find much of the economics literature on the political economy of reform naive or simplistic. However, what is encouraging about this literature is that economists are now doing their politicaleconomy analysis explicitly, rather than implicitly as used to be the case. Most economists have now come to the realization that good economic advice requires an understanding of the political economy of the situation. The result has been a remarkable degree of collaboration between economists and political scientists, as well as more work on political economy by younger economists. Both of these are good news. The bad news is that the habit of attributing myopia or irrationality to political actors— whether explicitly or, more often, implicitly—persists. As this essay attests, there is by now a wealth of case-study material on the experiences with policy reform. This mate-
jority better off. 37 Moreover, the same kind of argument explains why compensation, or the promise thereof, is not always an effective device to remove the distributional obstacle to reform. In the above example, ex ante losers know that if reform is passed, there will be an ex post majority in support of its continuation—even in the absence of compensation. Therefore, a promise to compensate losers ex post is not going to be credible. This kind of reasoning may explain why many reforms that would have been popular ex post are passed up ex ante. It may also explain why reforms that are instituted by an authoritarian regime against prevailing political sentiment survive the return of democracy (e.g., Pinochet’s reforms in Chile). Inspired more directly by the Eastern European context, Dewatripont and Roland (1992) analyze the case of an agenda-setting government faced with the task of shrinking the size of the state sector subject to political constraints. Some of their analysis focuses on the case where a reform must be approved by unanimity to be passed. Efficiency calls for immediate shrinkage of the state sector (i.e., shock therapy), as productivity is much lower than in the private sector. But unanimity requires that statesector workers who are forced to exit be compensated. Suppose that state-sector workers are heterogeneous in terms of their earning power in the private sector (to which they will be displaced) and that this information is private. The budgetary cost of shock therapy will be large, as all exiting workers will have to be paid a common exit fee that is based on the income loss that would have been incurred by the least-capable worker. If, in turn, raising fiscal revenue is distortionary, shock therapy may be dominated by par37 The situation is worse when individuals are risk averse. But the argument does not rely on risk aversion.
Rodrik: Understanding Economic Policy Reform rial provides plenty to chew on for anyone interested in the interactions between economics and politics. The next step should be to integrate evidence from these case studies more systematically into the analytical work, and in turn to use the analytical models as a springboard for more rigorous case studies. I close by highlighting one of the many questions raised earlier as requiring more work. Because distributional issues are at the heart of the literature discussed here, we need more progress on understanding why institutions for compensating losers from reform are not more common. There are very few papers where the difficulties of compensation are made endogenous to the analytical framework.38 This makes the literature somewhat incomplete in its diagnosis of the issues. It also opens up a natural avenue for future research.
REFERENCES ALESINA , A LBERTO AND D RAZEN, A LLAN. “Why Are Stabilizations Delayed?” Amer. Econ. Rev., Dec. 1991, 81(5), pp. 1170–88. ALESINA , A LBERTO AND R ODRIK , D ANI. “Distributive Politics and Economic Growth,” Quart. J. Econ., May 1994, 109(2), pp. 465–90. ALESINA , ALBERTO AND T ABELLINI , G UIDO . “External Debt, Capital Flight and Political Risk,” J. Int. Econ., Nov. 1989, 27(3/4), pp. 199–220. AMSDEN, A LICE H. Asia’s next giant: South Korea and late industrialization. New York and Oxford: Oxford U. Press, 1989. BALASSA , BELA AND ASSOCIATES . The structure of protection in developing countries. Baltimore: Johns Hopkins U. Press, 1971. BARRO , ROBERT AND WOLF , H OLGER . “Data Appendix for Economic Growth in a Cross-Section of Countries.” Unpub. paper, 1989. BATES , R OBERT H. AND KRUEGER , A NNE O. Political and economic interactions in economic policy reform. Oxford: Blackwell, 1993. BERG , A NDREW . “Measurement and Mismeasure38 Dewatripont and Roland (1992) and Fernandez and Rodrik (1991) are exceptions. In Dewatripont and Roland, the difficulty is created by the unobservability of individual abilities. In Fernandez and Rodrik (1991) compensation turns out to be time-inconsistent when it is known that reform will garner an ex post majority even absent compensation.
39
ment of Economic Activity During Transition to the Market” in Eastern Europe in transition: From recession to growth. Eds.: M ARIO I. B LEJER ET AL . Washington, DC: World Bank, 1993, pp. 39–63. B HAGWATI , J AGDISH . Foreign trade regimes and economic development. Vol. 11. Anatomy and consequences of exchange control regimes. Lexington, MA: Ballinger, 1978. ———. India’s economy: The shackled giant. Oxford, Clarendon Press, 1993. B IRDSALL , N ANCY; R OSS , D AVID AND S ABOT, R ICHARD . “Inequality and Growth Reconsidered.” Unpub. paper, Feb. 1994. B OSWORTH , BARRY P.; DORNBUSCH , R UDIGER AND L ABÁN , R AÚL , eds. The Chilean economy: Policy lessons and challenges. Washington, DC: Brookings Institution, 1994. B RESSER PEREIRA , L UIZ CARLOS ; M ARAVALL , J OSÉ M ARIA AND PRZEWORSKI, A DAM . Economic reforms in new democracies: A socialdemocratic approach. Cambridge: Cambridge U. Press, 1993. CALVO , G UILLERMO A. AND VÉGH , C ARLOS A. “Inflation Stabilization and Nominal Anchors,” Contemp. Econ. Policy, Apr. 1994, 21(2), pp. 35–45. CASELLA , A LESANDRA AND EICHENGREEN, B ARRY. “Can Foreign Aid Accelerate Stabilization?” Unpub. paper, 1994. CLARKE , GEORGE R. G. “More Evidence on Income Distribution and Growth.” U. of Rochester, unpublished paper, 1993. COLLINS, S USAN M. AND P ARK , WON-A M . “External Debt and Macroeconomic Performance in South Korea,” in Developing country debt and economic performance, Vol 3. Eds.: JEFFREY D. S ACHS AND SUSAN M. COLLINS . Chicago: U. of Chicago Press, 1989, pp. 153–369. CORBO , VITTORIO AND FISCHER, S TANLEY. “Structural Adjustment, Stabilization and Policy Reform: Domestic and International Finance,” in Handbook of development economics, Vol. III. Eds.: J. R. B EHRMAN AND T. N. S RINIVASAN . Amsterdam: North-Holland, 1995. CORBO , V ITTORIO AND R OJAS , PATRICIO . “World Bank-Supported Adjustment Programs: Country Performance and Effectiveness,” in Adjustment lending revisited: Policies to restore growth. Eds.: VITTORIO CORBO , S TANLEY FISCHER , AND S TEVEN B. W EBB . Washington, DC: World Bank, 1992. pp. 23–36. CUKIERMAN , A LEX; EDWARDS , S EBASTIAN AND T ABELLINI , GUIDO . “Seignorage and Political Instability,” Amer. Econ. Rev., June 1992, 82(3), pp. 537–55. D E LONG , J. B RADFORD AND S UMMERS , L AWRENCE H. “Equipment Investment and Economic Growth,” Quart. J. Econ., May 1991, 106(2), pp. 445–502. D EWATRIPONT , M ATHIAS AND R OLAND , GERARD . “Economic Reform and Dynamic Political Constraints,” Rev. Econ. Stud., Oct. 1992, 59(4), pp. 703–30.
40
Journal of Economic Literature, Vol. XXXIV (March 1996)
I NTERNATIONAL M ONETARY F UND . World economic outlook. Washington, DC: The Fund, 1994. J ONES, L EROY P. AND SA KONG , I L. Government, business, and entrepreneurship in economic development: The Korean case. Cambridge, MA: Harvard U. Press, 1980. KIM , J ONG -I L AND LAU , LAWRENCE J. “The Source of Economic Growth of the Newly-Industrializing Countries on the Pacific Rim.” Unpub. paper. Stanford U., Dec. 1992. KRUEGER, A NNE O. Foreign trade regimes and economic development. Vol. 10. Liberalization attempts and consequences. Lexington, MA: Ballinger, 1978. ———. Political economy of policy reform in developing countries. Cambridge, MA and London: The MIT Press, 1993. L ABÁN, R AÚL AND S TURZENEGGER , F EDERICO . “Distributional Conflict, Financial Adaptation and Delayed Stabilization,” Economics and Politics, Nov. 1994a, 6(3), pp. 257–76. ———. “Fiscal Conservatism as a Response to the Debt Crisis,” J. Devel. Econ., Dec. 1994b, 45(2), pp. 305–24. L AGO , R ICARDO. “The Illusion of Pursuing Redistribution through Macropolicy: Peru’s Heterodox Experience, 1985–1990,” in R UDIGER D ORNBUSCH AND S EBASTIAN EDWARDS 1991, pp. 263–323. L IN, CHING - YUAN. Industrialization in Taiwan, 1946–72: Trade and import-substitution policies for developing countries. New York: Praeger, 1973. L ITTLE , I AN M. D.; S CITOVSKY, T IBOR AND S COTT, M AURICE . Industry and trade in some developing countries. London: Oxford U. Press, 1970. L ITTLE , I AN M. D. ET AL . Boom, crisis, and adjustment: The macroeconomic experience of developing countries. New York: Oxford U. Press for the World Bank, 1993. MONDINO, G UILLERMO ; S TURZENEGGER , FEDERICO AND T OMMASI , M ARIANO . “Recurrent High Inflation and Stabilization: A Dynamic Game.” Unpub. paper, 1992. M OSLEY, PAUL . Conditionality as bargaining process: Structural-adjustment lending, 1980–86. Princeton, NJ: Princeton Essays in International Finance No. 168, 1987. Ö ZLER , SULE AND T ABELLINI , G UIDO . “External Debt and Political Instability.” National Bureau of Economic Research Working Paper, 1991. PANAGARIYA , ARVIND . “India: A New Tiger on the Block?” J. International Affairs, Summer 1994, 48(1), pp. 193–221. Penn World Table 5.5. A LAN H ESTON ET AL . Distributed by the National Bureau of Economic Research. Cambridge, MA, 1993. PEROTTI , R OBERTO . “Increasing Returns to Scale, Politics and the Timing of Stabilization.” Columbia U., unpublished paper, 1992a.
D ÍAZ-ALEJANDRO , CARLOS F. “Trade Policies and Economic Development,” in International trade and finance: Frontiers for research. Ed.: PETER B. KENEN . New York: Cambridge U. Press, 1975, pp. 93–150. D OLLAR , D AVID . “Outward-Oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976–1985,” Econ. Devel. Cult. Change, Apr. 1992, pp. 523–44. D ORNBUSCH , R UDIGER AND E DWARDS , SEBASTIAN , eds. The macroeconomics of populism in Latin America. Chicago and London: U. of Chicago Press, 1991. D ORNBUSCH , R UDIGER AND F ISCHER, S TANLEY. “Moderate Inflation,” World Bank Econ. Rev., Jan. 1993, 7(1), pp. 1–44. EASTERLY, WILLIAM . “How Much Do Distortions Affect Growth?” J. Monet. Econ., 1993, 32(2), pp. 187–212. EASTERLY, WILLIAM ; R ODRÍGUEZ , CARLOS A LFREDO AND S CHMIDT -H EBBEL , K LAUS . Public sector deficits and macroeconomic performance. New York: Oxford U. Press, 1994. EDWARDS, SEBASTIAN. “Trade and Industrial Policy Reform in Latin America.” NBER Working Paper No. 4772, June 1994. FERNANDEZ , R AQUEL AND R ODRIK , D ANI . “Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty,” Amer. Econ. Rev., Dec. 1991, 81(5), pp. 1146– 55. FISHLOW , ALBERT ET AL . Miracle or design? Lessons from East Asia. Washington, DC: Overseas Development Council, 1994. FLEMMING , JOHN. “Equitable Economic Transformation.” Paper presented to the Conference on Income Distribution and Sustainable Growth, IMF, Washington, DC, June 1995. GRINDLE , MERILEE S. AND T HOMAS , J OHN W. Public choices and policy change: The political economy of reform in developing countries. Baltimore and London: Johns Hopkins U. Press, 1991. GUITIAN, M ANUEL. “Economic Management and International Monetary Fund Conditionality,” in Adjustment and financing in the developing world. Ed.: T ONY KILLICK . Washington, DC: International Monetary Fund, 1982, pp. 73– 104. H AGGARD , STEPHAN AND KAUFMAN , R OBERT R., eds. The politics of economic adjustment. Princeton, NJ: Princeton U. Press, 1992. H AGGARD , STEPHAN AND WEBB, S TEVEN , eds. Voting for reform: Democracy, political liberalization, and economic adjustment. New York: Oxford U. Press, 1994. H ARBERGER, A RNOLD C. “Secrets of Success: A Handful of Heroes,” Amer. Econ. Rev., May 1993, 83(2), pp. 343–50. H ESTON , A LAN AND SUMMERS , R OBERT . “A New Set of International Comparisons of Real Product and Price Levels: Estimates for 130 Countries, 1950–1985,” Rev. Income Wealth, 1988, 34, pp. 1–25.
Rodrik: Understanding Economic Policy Reform
———. “Income Distribution, Politics, and Growth,” Amer. Econ. Rev., May 1992b, 80(2), pp. 311–16. PERSSON, T ORSTEN AND T ABELLINI , GUIDO . “Is Inequality Harmful for Growth?” Amer. Econ. Rev., June 1994, 84(3), pp. 600–21. PIÑERA , J OSÉ . “Chile,” in J OHN WILLIAMSON, ed. 1994. POLAK , JACQUES J. The changing nature of IMF conditionality. Princeton, NJ: Princeton Essays in International Finance No. 184, 1991. PRZEWORSKI , ADAM . Democracy and the market: Political and economic reforms in Eastern Europe and Latin America. Cambridge: Cambridge U. Press, 1991. R ANIS , GUSTAV AND M AHMOOD , S YED AKHTAR. The political economy of development policy change. Cambridge, MA and Oxford: Basil Blackwell, 1992. R ODRIK , D ANI . “King Kong Meets Godzilla: The World Bank and The East Asian Miracle,” (1994a) in ALBERT FISHLOW ET AL. 1994. ———. “The Rush to Free Trade in the Developing World: Why So Late? Why Now? Will It Last?” (1994b) in S TEPHAN H AGGARD AND S TEVEN B. WEBB, eds. 1994. ———. “Getting Interventions Right: How South Korea and Taiwan Grew Rich,” in Econ. Policy, Apr. 1995a, 20, pp. 55–107. ———. “Trade and Industrial Policy Reform,” in Handbook of development economics. Vol. IIIB. Eds.: J. R. B EHRMAN AND T. N. S RINIVASAN. Amsterdam: North-Holland, 1995b. ———. “The Dynamics of Political Support for Reform in Economies in Transition,” J. Japa-
41
nese Int. Econ., Special Issue on Transition, forthcoming. S ACHS , J EFFREY. “The Political Challenge of Economic Transition: The Case of Poland.” Unpub. paper. Jan. 1994. T AYLOR , L ANCE , ed. The rocky road to reform. Cambridge, MA: The MIT Press, 1993. T HOMAS , V INOD ET AL. Restructuring economies in distress: Policy reform and the World Bank. Oxford and New York, Oxford U. Press, 1991. T OMMASI , M ARIANO AND V ELASCO , A NDRÉS . “Where Are We in the Political Economy of Reform?” Unpub. paper, May 1995. V ELASCO , A NDRES . “Are Balance-of-Payments Crises Rational?” New York U., unpub. paper, 1994. WADE , R OBERT . Governing the market: Economic theory and the role of government in East Asian industrialization. Princeton, NJ: Princeton U. Press, 1990. WEI , S HANG -J IN. “Gradualism versus Big Bang: Speed and Sustainability of Reforms.” Unpub. paper. Harvard U., Dec. 1992. WILLIAMSON , J OHN, ed. The political economy of policy reform. Washington, DC: Institute for International Economics, 1994. WORLD BANK . The East Asian miracle. Washington, DC: World Bank, 1993. Y OUNG , A LWYN. “Lessons from the East Asian NICs: A Contrarian View.” NBER Working paper No. 4482. 1993. ———. “The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth Experience.” Unpub. paper. Feb., 1994.
Rodrik: Understanding ofEconomic Policy Reform Journal Economic Literature, Vol. XXXIV (March 1996)
Understanding Economic Policy Reform
By DANI RODRIK Columbia University
This paper has benefited from the comments of Herschel Grossman, Anne Krueger, Gustav Ranis, Jeffrey Sachs, John Williamson, and three referees.
I. Introduction
HIS GOVERNMENT will be austere, uncompromising, and unpopular if that is what is required to achieve economic recovery,” declared Mario Soares in 1983 upon taking office as prime minister of Portugal.1 This kind of talk is music to the ears of economists. Reform requires austere policies which respect budget constraints. It also precludes compromising with the narrow, special interest groups that have been the beneficiaries of the deleterious policies of the past. Policy makers who have courageously taken on such interest groups and pursued market-oriented policies are the heroes of the economics profession (see Arnold C. Harberger 1993). That such policies should also be unpopular (as Soares feared they would), however, requires a lot more explaining. What is the point of loudly proclaiming reforms if these are not aimed at improving the well-being of a large majority of the population? And if that is their goal, why should reforms be unpopular? In many areas of policy, there may ex1 Quoted by José Maria Maravall in Luiz Carlos Bresser Pereira, Maravall, and Adam Przeworski (1993).
“
T
ist “technical” uncertainty as to what the appropriate solution is to the problems at hand. Think of President Clinton’s health care plan, for example, or of global warming. Consequently, reforms will arouse opposition if they are viewed as applying the wrong fix or if they are perceived as being primarily redistributive (that is, zero-sum). What is remarkable about current fashions in economic development policy (as applied to both developing and transitional economies), however, is the extent of convergence that has developed on the broad outlines of what constitutes an appropriate economic strategy. This strategy emphasizes fiscal rectitude, competitive exchange rates, free trade, privatization, undistorted market prices, and limited intervention (save for encouraging exports, education, and infrastructure). Faith in the desirability and efficacy of these policies unites the vast majority of professional economists in the developed world who are concerned with issues of development.2
2 The convergence is not complete of course. But compared to two decades ago, the various sides have moved substantially closer to each other. One indicator of this is the recent book by Bresser Pereira, Maravall, and Przeworski (1993), which advocates a “social democratic” approach. The views expressed in this book concede an inor-
9
10
Journal of Economic Literature, Vol. XXXIV (March 1996) we observe such instances of collective irrationality. The events of the last decade have underscored the need to understand the political-economy of policy making. One of the eventual consequences of the global debt crisis that erupted in 1982 was a wave of market-oriented economic reforms, the likes of which have never been seen. The reforms were strongest and most sustained in Latin America, where countries like Bolivia, Mexico, Argentina, Peru, Colombia, and Brazil joined Chile in orthodoxy. But this was very much a global phenomenon. “Stabilization” and “structural adjustment” became the primary preoccupation of government leaders in Asia and Africa as well, even though the commitment to economic orthodoxy varied across countries and over time. These countries were in turn soon joined by the previously socialist economies of Eastern Europe and the former Soviet Union. Economists who had cut their teeth in Latin America’s economic quagmires became the advisors and analysts of these transitional economies. Even India, the giant archetype of a closed, import-substituting economy among developing countries, embarked on a process of economic liberalization in 1991 (see Jagdish Bhagwati 1993 and Arvind Panagariya 1994). These reforms were encouraging to economists and a vindication of sorts to those among them who had long advocated market-oriented reforms. But they in turn raise their own puzzles. Most fundamental of all, why are so many governments reforming now, after decades of adherence to policies of an opposite kind? This question poses a particularly important challenge to political economists: an understanding of these countries’ experiences now requires a theory that explains not only why seemingly dysfunctional policies had been initially un-
Hence economists are often torn between two conflicting perspectives: on the one hand, good economic policy should produce favorable outcomes and therefore should prove also to be good politics; on the other hand, the implementation of good economic policy is often viewed as requiring “strong” and “autonomous” (not to say authoritarian) leadership. The experience of Chile, a country which has perhaps gone further than any other in implementing liberal economic policies, provides a good example. An essay on Chile’s reform strategy by José Piñera (1994), an economist and minister of labor and social security under General Pinochet, concludes: “[i]n the end, good policy is good politics” (p. 231). The irony is that most of the reforms the author glowingly discusses in the preceding pages required the suspension of normal politics and as heavy a dose of authoritarianism as seen anywhere. Good economics does often turn out to be good politics, but only eventually. Policies that work do become popular, but the time lag can be long enough for the relationship not to be exploitable by would-be reformers. In Chile’s case, free market policies (implemented after 1973) were eventually resoundingly endorsed in the presidential elections of 1989 and have become the envy of Latin America.3 Conversely, bad economics can be popular, if only temporarily. President Alan García’s popularity soared in Peru during his first two years in office (1985–86), thanks to expansionary fiscal policies whose medium-term unsustainability should have been obvious to anyone with common sense (see Ricardo Lago 1991). The puzzle is why dinate amount to the consensus view, and depart from it in remarkably few details. I will discuss this book in Section IV. 3 For a recent evaluation, see Barry Bosworth, Rudiger Dornbusch, and Ral Labn (1994).
Rodrik: Understanding Economic Policy Reform dertaken and then maintained for so long, but also why these policies were suddenly abandoned en masse during the 1980s, often by the same politicians who had been among their most ardent supporters. Second, while the reforms were inspired at least in part by the East Asian experience, they took place much more quickly and, in many areas, are going considerably beyond those undertaken in East Asia. This raises the question of whether the new wave of reformers have internalized the correct lessons from the East Asian experience. Finally, are there any helpful rules for reformers to follow in guiding their policies through complicated political terrain? Can one hope to develop a “how-to” manual for the reformist politician? Puzzlement over such questions has led to a large and growing literature. A very short bibliography would include books by Merilee Grindle and John Thomas (1991), Robert Bates and Krueger (1993), Krueger (1993), Przeworski (1991), Ranis and Syed Mahmood (1992), Bresser Pereira, Maravall, and Przeworski (1993), Stephan Haggard and Robert Kaufman (1992), Dornbusch and Sebastian Edwards (1993), Haggard and Steven Webb (1994), Lance Taylor (1994), Williamson (1994), and Ian Little et al. (1993), not to mention countless papers. 4 As this partial list indicates, both economists and political scientists have devoted their attention to these issues, often together in coauthored or coedited works. Indeed, no other area of economics or political science that I can think of has spawned so much interdisciplinary work.5 In this essay, I will provide an econo4 One recent survey—Mariano Tommasi and Andres Velasco 1995—which overlaps with this one deserves special mention. 5 The literature on the economics of policy reform is of course even larger. For recent surveys, see Vittorio Corbo and Stanley Fischer (1995) and Rodrik (1995b).
11
mist’s perspective on the political economy of policy reform. I begin by examining the origins and analytical content of the new orthodoxy in development policy (Section II). I will focus here on two issues in particular which I feel remain in need of clarification. One of these concerns the distinction between (a) macroeconomic policies aimed at economic stability, such as fiscal, monetary, and exchange rate policies, and (b) liberalization policies aimed at structural reform and growth, such as the removal of relative-price distortions and the reduction of state intervention. It has become commonplace to conflate these two groups of policies, but for analytical purposes they are best kept apart. As we shall see, they also have different political-economy underpinnings. Moreover, maintaining the distinction reminds us that the consensus on what constitutes appropriate structural reform is based on much shakier theoretical and empirical grounds than is the consensus on the need for macroeconomic stability. The second issue concerns the appropriate lessons to be drawn from the experience of East Asian success stories. The new orthodoxy has tended to draw a somewhat biased picture that needs correction. Next, I will turn to the reforms of the 1980s and 1990s. This experience has opened an important window on the motivations of politicians, as well as on the nature of interactions between the economy and the polity. As indicated above, an important question is why so many countries have suddenly caught the reform bug. The confluence of economic crisis with reform has led to the natural supposition that crisis is the instigator of reform, a hypothesis that keeps reappearing in the literature and yet is inadequately analyzed. Section III discusses this issue, as well as related questions such as: what, if any, are the short-term
12
Journal of Economic Literature, Vol. XXXIV (March 1996) parently successful example of Soviet planning invited emulation. Moreover, the economic ideas of the 1950s and early 1960s tended to dismiss the benefits from trade and emphasized the need for physical capital accumulation and infant-industry promotion. As Krueger puts it,
[t]he underlying premises regarding markets and governments implicit in these policy prescriptions are obvious: There was a strong emphasis on the primacy of market imperfections. Market failures were thought to be relatively strong, while it was assumed that governments could correctly identify and perform economic functions. Virtually no attention was given to the possibility that there might be government failure. (p. 49)
costs of reform? and does foreign aid help or hamper reform? Finally, I will discuss some normative aspects related to the politics of policy reform (Section IV). In particular, I will analyze what kind of guidance, if any, the literature provides to the reformist politician regarding the appropriate strategies for carrying out his program. The focus here will be on the conflict between top-down versus participatory approaches. This discussion will take us back to some of the deeper questions about the collective rationality of political institutions and behavior. II. The New Orthodoxy in Development Thinking and Enduring Puzzles Once upon a time, there was something called the Third World and we thought we understood how it worked (or did not work). Countries of the Third World followed import-substitution policies, so called because the overarching objective of economic policy was to develop domestic manufacturing capability for goods previously imported. Such policies included import controls, overvalued exchange rates, binding ceilings on interest rates, a heavy dose of public ownership, and pervasive price regulation. The political-economy counterpart of these policies was the predominance of urban over rural interests, and within the urban sector, an uneasy alliance of sorts between the protected industries and the bureaucrats administering the protection. Perhaps no other economist has done as much as Anne Krueger in documenting and popularizing the shortcomings of import-substituting policies. As she points out in her Ohlin Lectures (1993), there was a multitude of reasons for the initial adoption of these policies. Newly independent governments had a strong desire for industrialization, and the ap-
And there was plenty of government failure. Not only did many infant industries fail to mature, but many countries succumbed to stop-go cycles driven by excessive government spending. Krueger explains that in viewing the government as a “benevolent social guardian” most economists had ignored a number of important forces at work. Individuals in the public sector were apt to follow their own selfish interests. They would be lobbied by pressure groups aiming to impose their own agenda on a largely docile majority. Policy interventions would create rent-seeking incentives diverting entrepreneurs from productive activities. Finally, the informational disadvantage of government bureaucracies over market participants would doom even the best-laid plans to inefficiency. Meanwhile, four East Asian economies were making a mockery of the export pessimism that had persuaded policy makers elsewhere to follow an inwardoriented strategy. South Korea and Taiwan, in particular, were able to engineer a remarkable increase in their growth rates, thanks to sharp jumps in their investment and export efforts during the
Rodrik: Understanding Economic Policy Reform
TABLE 1 COMPARATIVE GROWTH EXPERIENCE Per capita GDP, 1960 (1985 dollars) 883 1359 873 1017 1128 1745 2798 3294 Per capita GDP, 1989 (1985 dollars) 6206 8207 815 1082 756 4138 5163 3608
13
Country South Korea Taiwan Ghana Senegal Mozambique Brazil Mexico Argentina
Per capita GDP growth, 1960–89 (%) 6.82 6.17 −0.54 0.16 −2.29 3.58 2.36 0.63
Source: Penn World Table 5.5 (1993).
mid-1960s. Table 1 summarizes this impressive performance in comparative context. Recent work by Jong-Il Kim and Lawrence J. Lau (1992) and Alwyn Young (1993, 1994) has shown that these were miracles of accumulation rather than of productivity: the sharp increases in physical and human capital as well as in labor-force participation account for virtually all of the rise in output, and consequently the East Asian tigers’ performance with respect to total factor productivity (TFP) growth does not look outstanding. Of course, for accumulation to have taken place at such rates, the profitability of investment in the region must have been very high, which needs explanation. What was the key to these economies’ success? Among professional economists, there soon developed the view that the East Asian miracles could be attributed to market-oriented policies and the reduced role of government intervention. Hence, Ranis and Mahmood (1992, p. 138) attribute the Taiwanese and South Korean successes to the willingness of the governments in both countries—albeit somewhat less so in South Korea—to allow growth to proceed along a “natural” path and an aversion to use covert
measures of resource transfers in order to promote growth artificially.
And Krueger (1993, p. 30) writes:
To be sure, the Korean economy has not been characterized and is not characterized by laissez-faire. But in contrast to the overcontrolled, overregulated, highly distorted economies described above, the Korean economy has been characterized by diminishing intervention in most spheres of economic activity, and the degree of distortion is considerably smaller.
Indeed, in both Korea and Taiwan, there was an extensive set of reforms during the late 1950s and early 1960s, on which more later. So the new orthodoxy was built on two mutually reinforcing pillars: one was the set of policies that had been tried by the import-substituting countries and had failed; the second was the set of successful policies implemented by the East Asian tigers. These two sets of policies bear close scrutiny, as they overlap to a much greater extent than the orthodox case likes to admit. A. Macroeconomic Disequilibrium or Import-Substitution Policies? In describing the experience of developing countries it has become common
14
Journal of Economic Literature, Vol. XXXIV (March 1996) rates are different. By definition, they result in trade deficits that are unsustainably large, and therefore in balanceof-payments crises (unless they are corrected quickly). Consequently, exchange rate misalignments are closely associated with economic instability and with the deterioration of economic performance over the medium to long run. The confusion between these two types of policies often reveals itself in empirical studies that uncover a negative relationship between measures of exchangerate distortion and economic growth, and then attribute the effect to the lack of openness in the sense of trade protection. 7 Similarly, consider subsidies to specific industries versus large fiscal deficits. The former can be damaging if the targeted industries do not produce positive externalities or if the subsidies lead to rent-seeking activities. But the damage done by large, sustained budget deficits is often much larger. An unsustainable deficit can be financed by foreign borrowing or by monetization, and in either case a crisis is often not too far away. 8 Usually, external sources are the first resort, and an inflation problem follows on the heels of a payments crisis. The World Bank recently undertook a large-scale research project on the consequences of public-sector deficits in developing countries. The findings were unequivocal: “Deficits . . . are unambiguously bad for growth.”9 This distinc7 See for example the use made of David Dollar’s (1992) “real exchange rate distortion” index in World Bank (1993). 8 Of course, foreign borrowing or moderate amounts of monetization (seignorage or inflation tax) need not be bad things. The focus here is on unsustainable levels of borrowing or seignorage. 9 William Easterly, Carlos Alfredo Rodrguez, and Klaus Schmidt-Hebbel (1994, p. 1). It is of course possible, as a referee pointed out, that a deeper underlying cause (i.e., political instability) is jointly responsible for both the deficits and low growth.
to lump together a wide range of policies, as I too have done above, under the label of “import-substitution policies.” For descriptive purposes, this makes perfect sense. Except for a handful of countries in East Asia, most developing countries did combine illiberal trade and price policies with (at least occasional) fiscal profligacy and overvalued exchange rates. So there was a common syndrome, which perhaps does deserve to go under a single name. However, the practice has also frequently led economic analysis astray and generated confusion. The trouble is that failures were often misattributed to microeconomic policies, when their sources lay either with unsustainable macroeconomic policies or bureaucratic and institutional shortcomings. Consider for example two trade-related policies: import restrictions (e.g., import quotas) and overvalued exchange rates. These two have rather different implications for economic stability and long-run performance. The effect of import restrictions is to repress trade (imports as well as exports in the longer run). While this is costly insofar as it engenders some resource misallocation, it does not inherently generate economic instability, nor does it necessarily reduce long-term growth.6 Overvalued exchange
6 In traditional economic theory, trade restrictions have level effects, but no growth effects. That is, a 20 percent tariff may reduce real income by, say, 0.5 percent of GDP (permanently), but it will not affect the economy’s long run growth rate. In the more recent endogenous growth literature, trade restrictions may have growth effects, but these growth effects can go either way depending on the model. For a small economy, the effect tends to depend on whether the comparativeadvantage effect pulls resources in or out of the growth-sectors of the economy (those with technological externalities or learning-by-doing). Empirical work on growth has often claimed to discover a negative relationship between trade protection and growth. However, this literature is marred by severe analytical and conceptual confusions, with macroeconomic policies often confused for trade restrictions and the endogeneity of trade policy not fully accounted for.
Rodrik: Understanding Economic Policy Reform
TABLE 2 DETERMINANTS OF THE DEBT CRISIS, 1982 Failure to adjust monetary and fiscal policy Yes Yes Yes Yes Yes Yes Yes Yes
15
Large external shock Troubled Counties Argentina Brazil Chile Costa Rica Cote d’lvoire Mexico Morocco Nigeria Unweighted Average Moderately Troubled Countries Colombia Kenya Sri Lanka Unweighted Average Untroubled Countries Cameroon India Indonesia Korea Pakistan Thailand Turkey Unweighted Average No Yes Yes No Yes No No No
Index of relative-price distortion 0.3054 0.2019 0.4460 0.2818 0.2438 n.a. 0.2675 0.2306 0.2824 0.2744 0.1218 0.8606 0.4189 0.2344 0.2620 0.4503 0.2128 0.3814 n.a. n.a. 0.3082
No Yes Yes
Yes Yes No
Yes No No Yes No Yes No
No No No No No No No
Source: Little et al. (1993), Table 4.4, except for the relative-price distortion index which is taken from Easterly (1993). The latter index is the variance of the log input prices (relative to U.S. prices) across commodities, measured in 1980. See Easterly (1993) for the method of calculation and the justification for the index.
tion—between micro and macro—is lost in many studies that attribute balanceof-payments crises to activist industrial policies (rather than the unwillingness or inability to balance budgets). The watershed event of the 1980s for most developing countries, the generalized debt crisis that followed the Mexican moratorium of August 1982, was a dramatic confirmation of the importance of prudent macroeconomic policies. Looking below the surface, it was evident that the crisis affected only those countries that did not respect budget constraints. India, the import-substitut-
ing country par excellence, managed to escape the debt crisis during the 1980s, thanks to its tradition of conservative monetary and fiscal policies.10 At the other end of the spectrum, South Korea experienced a payments crisis in 1979– 80, before the Latin American countries, as a consequence of an ambitious investment program running ahead of available domestic savings. But the quick adoption
10 During the 1980s, however, a growing publicsector deficit began to endanger macroeconomic balances. In 1990, the deficit reached 8.3 percent of GDP, culminating in a balance-of-payments crisis in 1991 (Panagariya 1994, pp. 205–06).
16
Journal of Economic Literature, Vol. XXXIV (March 1996) most everything right on the microeconomic front, but still face a deep crisis if macro policies—in Mexico’s case an overvalued exchange rate aggravated by pre-election credit expansion—are not managed well. This distinction between microeconomic distortions and macroeconomic stability is one that economists have long recognized. Yet it is also one that has made little impression on the development profession. In an essay published in 1975, for example, Carlos Díaz-Alejandro (1975) faulted the Little-Scitovsky-Scott (1970) study mentioned above for lumping together all features of the import-substitution sydrome, such as import and other controls, tariffs, overvalued and pegged exchange rates, spectacular balance-of-payments crises, inflationary pressures, and stop-go cycles. Following a “guilt by association” procedure, they then tend to blame much of what is going wrong in less developed countries on that ill-defined syndrome. Unsophisticated readers may indeed conclude that nearly everything gone wrong in those countries is due to that wicked syndrome. Consider a mental experiment. What would have happened if, say, Argentina and Colombia had adopted flexible exchange rates back in 1945, while adopting also an across-theboard import tariff of 150 per cent ad valorem? I suspect their record, at least on growth and exports, would have been much better. Their harmful stop-go policies may be blamed to a large extent on exchange-rate mismanagement . . . and on other short-run policies that could be analytically separated from the long-run effects of protection. (Díaz-Alejandro 1975, pp. 115–16, emphasis in the original)
of expenditure-reducing and expenditure-switching policies enabled a swift recovery, after a five percent contraction in 1980 (see Susan Collins and Won-Am Park 1989). Before long, Korea’s crisis was forgotten. More broadly, there was no correlation between the propensity to fall into crisis and the nature of microeconomic policies proper. As Table 2 indicates, the countries that experienced a debt crisis in 1982 were those that failed to adjust their monetary and fiscal policies, and not those that had large microeconomic distortions or were confronted with particularly large external shocks. There is little question that the microeconomic distortions in themselves were often costly. A series of team efforts led by Little, Tibor Scitovsky, and Maurice Scott (1970), Bela Balassa (1971), Bhagwati (1978), and Krueger (1978) had documented in detail the resource-allocation costs of high protection. These studies used measures of effective rates of protection (ERP) and domestic resource costs (DRC) to demonstrate that the governments’ price and trade policies had led to the creation of wildly inefficient industries, at times producing negative value-added at world prices! Moreover, the net incentive effects of these policies seemed haphazard, judged by the variability of ERPs and DRCs across industries. More recent work has shown that particular relative price distortions, those affecting capital and intermediate goods, can be costly to growth as well (Bradford De Long and Lawrence Summers 1991; Easterly 1993). However, what eventually drove many import-substituting countries to ruin were not such micro- economic inefficiencies, but macroeconomic imbalances and the inability to correct them with sufficient speed. Mexico’s currrency crisis in December 1994–January 1995 is a poignant example of how one can get
Elementary as these points may be, they were largely overlooked in the aftermath of the debt crisis. The consensus post mortem view held the whole complex of import-substitution policies responsible for what was essentially a crisis of overspending exacerbated by the fickleness of international capital markets. It became commonplace to view the debt
Rodrik: Understanding Economic Policy Reform
TABLE 3 THE “WASHINGTON CONSENSUS” AND EAST ASIA Elements of the Washington Consensus 1. Fiscal discipline 2. Redirection of public expenditure priorities towards health, education and infrastructure 3. Tax reform, including the broadening of the tax base and cutting marginal tax rates 4. Unified and competitive exchange rates 5. Secure property rights
17
South Korea Yes, generally Yes Yes Yes
Taiwan
Yes, generally
Yes
6. Deregulation 7. Trade liberalization 8. Privatization
Yes (except for limited time periods) President Park starts his rule in 1961 by imprisoning leading businessmen and threatening confiscation of their assets. Limited Limited until the 1980s No. Government established many public enterprises during 1950s and 1960s. DFI heavily restricted Limited until the 1980s
Yes Yes
9. Elimination of barriers to direct foreign investment (DFI) 10. Financial liberalization.
Limited Limited until the 1980s No. Government established many public enterprises during 1950s and 1960s DFI subject to government control Limited until the 1980s
Source: Williamson (1994) for first column, and author’s evaluation.
crisis as the consequence of import-substitution (“inward-oriented”) policies. 11 The intellectual ground was therefore cleared for the wholesale reform of prevailing policies in Latin America, Africa, and Asia. Orthodox economists who had the ear of policy makers now had their chance to wipe the slate clean and mount a frontal attack on the entire range of policies in use. After some de11 This view was (and is) strongly held in the World Bank, which shaped development policy in many countries during the 1980s through its structural adjustment loans. See for example the World Bank volume edited by Vinod Thomas et al. (1991), and especially the essays by the editors and by Ernest Stern.
lay, this produced dramatic results, especially in Latin America. B. What Did the East Asian Governments Do Right? Ironically, many governments (notably in Latin America) ended up implementing policies that went far beyond what the East Asian governments themselves had adopted since the 1960s. As intimated above, many of the discredited policies had long been in use in South Korea, Taiwan, and Singapore, and apparently to good effect. (Hong Kong’s policies have come closest to the laissezfaire ideal, even though there is plenty of
18
Journal of Economic Literature, Vol. XXXIV (March 1996) fered in the investment decisions of private enterprises. And far from privatizing public enterprises, both countries actually increased their reliance on such enterprises during the crucial decade of the 1960s. 13 In short, where South Korea and Taiwan followed the orthodox path most closely was in maintaining conservative fiscal policies and competitive exchange rates; this accounts for their ability to avoid protracted periods of macroeconomic instability, particularly in the crisis-ridden decade of the 1980s. In the area of microeconomic interventions, however, their experience diverged from the orthodox path. By contrast, it is striking how many Latin American countries have come within reaching distance of completing the items on the “Washington consensus” in a period of no more than a few years during the 1980s. Mexico, Bolivia, and Argentina, to cite some of the more distinguished examples, have undertaken more trade and financial liberalization and privatization within five years than the East Asian countries have managed in three decades. I will discuss the reforms of the 1980s at greater length later. First we have to confront two puzzles that arise from the contrasting experiences of the East Asian tigers and other developing countries prior to the 1980s. How could the East Asian countries avoid the disasters that accompanied interventionist policies elsewhere? And why did most developing countries succumb so easily, and often periodically, to unsustainable fiscal and exchange-rate policies? Neither of the puzzles is handled very well in the existing literature, although we do have a number of useful leads.
13 Probably the two best known works on these issues are Alice Amsden (1989) and Robert Wade (1990). Leroy Jones and Il Sakong (1980) and Ching-yuan Lin (1973) are two older books which are extremely detailed and informative.
intervention in the housing market.) To appreciate this point, it is useful to spell out in somewhat greater detail the elements of the new orthodoxy, so we can compare them to East Asian policies. Fortunately, we have a useful list of policy desiderata compiled by Williamson, who has dubbed it the “Washington consensus.” Williamson originally compiled the list in 1990, and a summary is included in Williamson (1994). In Williamson’s words, “[t]he ‘Washington consensus’ offers a description of what is agreed about the set of measures that are typically called for in the first stage of policy reform. . . .” (p. 17). 12 The list is shown in Table 3, along with my own summary comments on the degree of compliance exhibited by Taiwan and South Korea in each area of reform. How well did South Korea and Taiwan do according to this list? Judging by the number of the prescriptions these countries did or did not follow, we would have to award South Korea a score of about five (out of ten), and Taiwan about six. Both countries managed fiscal expenditures and revenues rather well, avoiding macroeconomic stop-go cycles and high inflation. They also consistently maintained unified exchange rates (Taiwan since 1961 and Korea since 1964) and competitive parities for the most part. But the rest of the scorecard is less spectacular. Taiwan welcomed DFI, but Korea much less so. Korea repressed interest rates and made heavy use of subsidized credit; Taiwan did not do so, but did give priority to public enterprises in credit allocation. Neither country significantly liberalized its import regime until the 1980s. Both countries heavily inter12 Williamson is quick to say that there remains a number of disagreements among economists on some of the narrower issues, for example the speed of trade liberalization or the need to eliminate indexation. See pages 17–18 for his list of disagreements.
Rodrik: Understanding Economic Policy Reform
TABLE 4 HUMAN CAPITAL INDICATORS IN EAST ASIA: ACTUAL VERSUS PREDICTED VALUES FOR THE EARLY 1960s Primary enrollment ratio, 1960 Hong Kong Indonesia Japan Korea Malaysia Singapore Taiwan Thailand Act. Pred. Diff 0.87 0.83 0.04 0.67 0.51 0.16 1.03 0.92 0.11 0.94 0.57 0.37 0.96 0.68 0.28 1.11 0.78 0.33 0.96 0.62 0.34 0.83 0.57 0.26 Secondary enrollment ratio, 1960 Act. Pred. Diff 0.24 0.23 0.01 0.06 0.07 −0.01 0.74 0.29 0.45 0.27 0.10 0.17 0.19 0.15 0.04 0.32 0.21 0.11 0.28 0.12 0.15 0.12 0.10 0.02
19
Literacy rate, 1960 Act. 0.70 0.39 0.98 0.71 0.53 0.50 0.54 0.68 Pred. Diff 0.59 0.11 0.25 0.15 0.70 0.28 0.31 0.40 0.43 0.10 0.54 −0.04 0.36 0.18 0.31 0.37
Source: Authors’ calculations. Note: Predicted values of the indicators are obtained from a cross-country regression run on a 118-country sample, with per capita GDP in 1960 and its square used as independent variables. Source of the data: Alan Heston and Robert Summers (1988) and Robert Barro and Hoger Wolf (1989).
C. How Did East Asian Countries Manage to Intervene without Inviting Rent Seeking? One major puzzle noted above had to do with the apparently successful pursuit of interventionist microeconomic policies by South Korea and Taiwan. Why did trade protection, industrial policy, and subsidized credit work in these countries when it failed most everywhere else? There has been much debate about these questions, not the least about the degree of intervention itself. The extent to which activist government policies enabled these miracles of accumulation to take place is an issue that is not yet settled. 14 From our present perspective, we can ignore purely economic aspects and narrow down the question to the following: how were the East Asian governments able to avoid the rent-seeking activities that typically accompanied microeconomic interventions? While we have some clues, the answer is that we do not really know. There was
14 For two recent contrasting viewpoints, see World Bank (1993) and Albert Fishlow and others (1994).
clearly something special about the ability of the Taiwanese and South Korean policy makers to discipline their private sectors and their bureaucracies—an ability to which the label of “strong” or “hard” state is often attached. But where this ability came from, and whether it can be replicated in other settings, remains a mystery. Authoritarianism may have had something to do with the East Asian governance style, but there are too many mismanaged dictatorships around the world to take the hypothesis seriously. One need only look at Sub-Saharan Africa. Krueger is clearly thinking of the East Asian cases when she writes:
The adoption of the same economic policies in response to the same (economic) circumstances will . . . have different consequences under a politically strong leadership of a government with a well functioning bureaucracy capable of carrying out the wishes of the leadership than it will when a weak leadership of a coalition attempts to do the same things in circumstances where bureaucrats believe that they can generate support for opposition to those policies . . . (Krueger 1993, p. 9)
20
Journal of Economic Literature, Vol. XXXIV (March 1996) exceptionally equal by cross-country standards (Table 5). Equality may have been conducive to better governance for at least three different reasons. First, these governments did not generally have to contend with powerful industrial or landed interest groups: therefore, policy making and implementation could be insulated from pressure-group politics. Second, the absence of largescale inequities meant that governments felt no immediate need to undertake redistributive policies; they could concentrate on expanding the pie instead. Note that this point is analytically distinct from the previous one. Even a government which is completely insulated from lobbying groups would wish to redistribute income as long as its objective function looks anything like a conventional social welfare function. This redistribu-
But we do not get much more help than this acknowledgement that governments and bureaucracies differ. It is reasonable to suppose that at least part of the explanation has to do with with some of the special initial conditions that the East Asian countries shared prior to their economic take-off. Two such conditions stand out. First, by the late 1950s the East Asian economies had for the most part a much better educated labor force than would have been expected on the basis of their income levels (Table 4). This may have made it easier to establish a competent bureaucracy (as well as enhancing the productivity of interventions aimed at boosting private investment, as argued in Rodrik 1995a). Second, and perhaps more importantly, in all of them the distribution of income and wealth around 1960 was
TABLE 5 DISTRIBUTIONAL INDICATORS FOR EAST ASIAN AND COMPARATOR COUNTRIES, AROUND 1960 Gini Coefficient for Income, c. 1960 0.49 0.33 0.40 0.34 0.42 0.31 0.40 0.41 0.39 0.44 0.53 0.42 0.42 0.64 0.53 0.45 0.56 0.50 Gini Coefficient for Land Ownership, c. 1960 n.a. n.a. 0.47 0.39 0.47 0.46 n.a. 0.46 0.45 0.87 0.85 0.67 0.52 0.69 0.69 0.53 0.59 0.68
Country East Asia Hong Kong Indonesia Japan Korea Malaysia Taiwan Singapore Thailand Unweighted Average Others Argentina Brazil Egypt India Kenya Mexico Philippines Turkey Unweighted Average Source: Rodrik (1994a)
Rodrik: Understanding Economic Policy Reform tive motive figures more heavily the worse the pre-tax distribution of resources among the population.15 Third, and related to these, the fact that the top political leadership was free to focus on economic goals meant that it could supervise the bureaucracy closely and make sure that the bureaucrats assisted rather than hindered private entrepreneurship. For bureaucracies are prone to two problems that are fatal to economic performance: they can be captured by the interests they are supposed to regulate, and they can create excessive red tape discouraging economic activity. In Korea and Taiwan, these problems were avoided because the bureaucracies were very closely supervised by the top political leadership. These points have recently received some indirect empirical support in a number of papers which have found a positive relationship between equality and subsequent economic growth. See in particular Alesina and Rodrik (1994), Persson and Tabellini (1994), George Clarke (1993), and Nancy Birdsall, David Ross, and Richard Sabot (1994). In these papers, the initial level of income equality around 1960 is shown to be robustly and positively correlated with growth over the next three decades, controlling for other initial conditions such as per capita income and educational attainment. The theoretical models proposed by Alesina and Rodrik (1994) and Persson and Tabellini (1994) to explain this phenomenon rely on a political-economy argument. When distributive policy is sensitive to the preferences of the median voter, it can be shown that the equilibrium level of redistribution is increasing in the gap between the median voter’s income and average income.
15 The analytics backing up this statement can be found in Alberto Alesina and Rodrik (1994) and Torsten Persson and Guido Tabellini (1994). The result is explained later in the text.
21
More equality usually (but not necessarily) goes with a lower difference between median and average incomes. Consequently, and assuming that redistributive policies act as a tax on accumulation, societies with lower inequality will resort to less redistribution and grow faster. The empirical literature to date has not been successful in testing directly for the presumed political channel through which equality affects growth. Therefore the precise links between equality and growth remain to be demonstrated (Roberto Perotti 1992b). As the East Asian cases suggest, it is at least plausible that governance plays some role in this relationship. Nonetheless, this too is an area in need of more research. We need to understand better why broadly similar policies can produce quite different outcomes in different societies. D. The Political Economy of Macroeconomic Cycles The continued existence of some of the policies that make up the “import substitution” syndrome, despite mounting evidence of their inefficiencies, can be understood in distributional terms: such policies redistribute income or rents to favored groups in society. Their beneficiaries—business and labor groups protected from foreign and domestic competition—are naturally resistant to their reform. But what is distinctive about large-scale deficit spending and overvalued currencies is that these policies are by their very nature temporary, and the longer they are pursued the more drastic their eventual reversal must be. Groups that benefit during the upswing of the joyride have to suffer losses during the downswing. Labor, for example, normally gains from expansionary “populist” policies, but also ends up as the greatest loser when the eventual crisis takes its toll on real wages and employment. Businessmen benefit from
22
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 6 CONSEQUENCES OF POPULISM IN PERU Average 1980–84
1985 5.8 0.3 2.4 2.3 158.3 64
1986 9.1 6.0 9.5 13.3 62.9 73
1987 12.9 7.2 7.8 8.3 114.5 79
1988 15.6 7.4 −8.8 −11.5 1722.3 60
1989 9.8 1.0 −10.4 −7.5 2775.3 29
Public sector deficit (% of GDP) Current account deficit (% of GDP) Real GDP growth Real consumption growth Inflation Real wages (1979=100) Source: Lago (1991, Table 9.4).
7.8 3.9 −1.0 −0.4 87.0 95
cheap imported inputs while the currency overvaluation lasts, but they are condemned to take a hit when the inevitable devaluation takes place (or when a large black-market premium makes its appearance). Therefore, short of attributing myopia or irrationality to the main political actors, it is hard to understand why such policies find support in the first place. 16 A fascinating case in point is the experience of Peru under President García (1985–90). Facing a stagnant economy, the new García administration launched a “heterodox program” in August 1985. The main thrust of the program was to boost consumption demand by increasing real wages, subsidizing consumption, and creating public-works programs. The government also instituted a freeze on
16 Political scientists too often ignore this issue, by treating all policies as benefiting some groups necessarily at the expense of others. Therefore the presence or absence of reform is viewed in purely partisan terms. This view is taken to its logical extreme by Bates in his comment on Williamson’s (1994) introduction. He explains reform thus: “Economic technocrats become powerful, and thus reform becomes politically sustainable, when they serve the interests of powerful groups: industries, sectors, or regions of the economy” (in Williamson 1994, p. 32). The fact is that reform— particularly in the macroeconomic area—often benefits the politically influential groups that block it.
prices, interest rates, and the exchange rate. And consumption did boom for a couple of years, raising output alongside it (Table 6). According to Lago, “[p]rior to 1987, private-sector confidence in and support of the government’s economic policy could only be described as being unanimous” (1991, p. 281). However, the process was clearly unsustainable: the public-sector and current-account deficits both rose substantially and foreign reserves were depleted. By late 1988, the economy had collapsed and prices were near hyperinflation levels. Real wages, which had increased until 1988, took a sharp nose dive and in 1989 stood at a third of their 1987 level (Table 6). (For the full account, see Lago 1991.) Peru’s case may be extreme, but it is by no means unique. Observers of the developing world have long been fascinated by the prevalence of such boomand-bust cycles. Ranis and Mahmood (1992) document how governments embark on unsustainable spending booms during periods when resources are temporarily plentiful (due either to improvements in the terms of trade or the availability of external finance), and then resist retrenchment when the inevitable downturn arrives. What is the explanation for these cycles? The answers pro-
Rodrik: Understanding Economic Policy Reform vided in the case-study literature, as in Ranis and Mahmood, tend to be mechanical. Ranis and Mahmood (1992, p. vi) summarize their conception of political economy thus: we accept the concept that the typical LDC state is an instrument of the most powerful— if often myopic—interest groups, the new industrialist class, public or private.
23
The operative term here, for our purposes, is “myopic.” Ranis and Mahmood do not offer an explanation for these policy cycles, save for suggesting shortsighted behavior. At least these authors are explicit about their assumption of myopia. More commonly the attribution of myopia to policy makers and interest groups is implicit. For example, Krueger (1993, p. 19) writes:
Regardless of the form of government—benevolent guardian, predatory authoritarian or bureaucratic, or factional—the political process typically demanded more resources than were available from tax revenues in the early stages of growth.
She then lays out in great detail the corrosive and cumulative long-run consequences of deficit spending and fixed exchange rates. But she considers these consequences as largely unanticipated by the relevant groups. But myopic behavior can take us only so far in understanding these policy cycles. For one thing, the underlying economics is straightforward: it does not require a Ph.D. in economics to realize that spending money one does not have can result in unpleasant consequences.17 For another, people ought to learn from their mistakes. To take one—not altogether extreme—example, since the 1950s Turkey has gone through four full boom-and-bust cycles, one per decade.
17 Of course, that does not prevent some people from trying it anyhow. But we’d like to think of this as the exception rather than the rule.
Because economists hate to give up on rationality (at least until they become older—wiser?—and distinguished), it is not surprising that a small analytical literature has developed around this set of issues. Several recent papers have proposed formal models that generate political outcomes that are inefficient from the standpoint of the politically powerful groups themselves, even though these groups behave rationally and non-myopically. The trick is usually performed by positing some kind of coordination problem among the contending actors. Some of these models focus on providing an explanation for the adoption of unsustainable policies, while some focus on the timing of reform (usually fiscal stabilization). A few attempt to do both. In the near-classic of this field, Alesina and Allan Drazen (1991) show how a stabilization can be delayed, at great cost to the parties involved, thanks to a “war of attrition” between two groups, each of which is uncertain about the costs being incurred by the other group. It is individually rational to wait in this model because the group that caves in first is assumed to bear a larger part of the post-stabilization tax burden. Each group has the incentive to wait and see if the other group will throw in the towel first. Stabilization takes place only when one of the groups figures that it stands to gain more from assuming the cost of stabilization than from waiting another instant to see if its rival will do so instead. In other papers, Alesina and Tabellini (1989), Sule Özler and Tabellini (1991), and Alex Cukierman, Edwards, and Tabellini (1992) have shown how macroeconomic problems can arise from political instability (i.e., expected turnover in governments) as well as polarization. In these papers, governmental “myopia” arises not from a war of attrition, but from the realization by groups in power that they may be replaced by future gov-
24
Journal of Economic Literature, Vol. XXXIV (March 1996) oped by Velasco (1994), who considers a model in which two organized groups treat the state’s resources as a common pool, and decide every period how much to extract from the pool. For simplicity and tractability, the choice is limited to one of two options: One option is to extract a “large” share, which leads to a build-up of government debt and a reduction over time in the state’s net wealth. The other is to extract a “small” share, which leaves net government wealth unchanged. Velasco demonstrates the existence of a switching equilibrium in which the two groups demand large transfers until the public debt stock reaches a critical amount, after which they both switch to the more moderate strategy. The latter is a cooperative outcome maintained by trigger strategies, and stands for the stabilization stage. The intuition is that once government net wealth becomes very low (i.e., once the state’s resources are sufficiently plundered), the benefits of the defection strategy become low compared to its costs (which are modeled here as being constant and independent of net wealth). Finally, Perotti (1992a) builds a model with three groups—the poor, the middle class, and the capitalists—to show how unsustainable policies can emerge as a consequence of a “populist” alliance between the poor and the capitalists. Three key features lead to this result. First, it is assumed that capitalists can insulate themselves from the costs of stabilization in the long run because capital is mobile in the long run (and only in the long run). Second, the poor are too poor to be taxed, so they do not bear any of the costs of stabilization. Third, the poor derive a positive externality from the profits made by capitalists at home. Now suppose that the economy is hit with a negative shock which makes longrun capital flight the dominant strategy
ernments with different ideological or redistributive preferences. More recent papers by Labán and Federico Sturzenegger (1994a and 1994b) generate results with the same flavor, by relying not on asymmetric information or uncertainty but on the dynamics of the flight from the domestic currency once inflation sets in. These papers assume two contending social groups—the rich and the poor—which prefer, at low levels of inflation, not to submit themselves to a costly stabilization. The key dynamic is generated by the presence of a “financial adaptation” technology (i.e., bank accounts in Miami), which is available only to the rich (at a fixed cost), and which leads to both increased inflation and a greater burden on the poor over time. The main result is that at some point the poor may therefore become willing to accept conditions they would have earlier rejected. Guillermo Mondino, Sturzenegger, and Tommasi (1992) generalize this framework to generate the possibility of recurring cycles of inflation and stabilization. In this model, two political groups interact strategically and demand transfers from the government, with the transfers paid for by the inflation tax. In addition, members of each of these groups have access to financial adaptation, and they behave atomistically in deciding when to indulge in it. Because the demand for transfers is determined at the group level while the flight from domestic currency is decided at the individual level, socially suboptimal outcomes and cycles both become a possibility. Under low inflation it may be optimal for groups to demand high levels of transfers, which sets into motion rising levels of inflation through the flight from local currency. In turn, at high levels of inflation, it may be optimal for the groups to agree on a stabilization. A variant along the same lines is devel-
Rodrik: Understanding Economic Policy Reform for capitalists. Because domestic capital will be abroad in the long run anyhow and the burden of stabilization does not fall on the poor, the latter’s well-being is maximized by expansionist policies that increase domestic profits in the short run. Hence the poor and the capitalists ally themselves against the middle class in pursuit of unsustainable policies that they know will be reversed eventually. In these models, political agents or groups are assumed to be rational and forward-looking, with expectations that are consistent with the properties of the underlying model. Behavioral rules are derived from solving optimization problems with well-defined objective functions. The political-economic outcome is derived as a Nash equilibrium in which each individual or group is doing the best it can given the actions of others. Of course, none of these models is capable of providing the full political-economic story of the boom-and-bust cycles discussed above. But they each capture an aspect of the phenomenon, and perhaps more importantly, they confirm that we can do better than resort to myopia or irrationality when explaining social phenomena. There are several directions in which this literature could be usefully extended. First, we now need theoretically informed case studies—or more formal test—that attempt to discriminate among these alternative stories. Second, the normative implications of these models for policy and institutional design have to be worked out. Third, there are several relevant aspects of reality which have yet to receive attention. For example, in many political organizations (parties, unions, etc.) principal-agent problems prevent the leadership from internalizing in full the interests of the rank-and-file. It is plausible that this would be one source of inefficiency in
25
economic policy making.18 Similarly, there is no room in these models for learning about the way the economy functions: if it is people that learn rather than organizations, each generation will repeat the mistakes of the previous ones.19 Finally, and most fundamentally, these papers leave hanging a key question: if distributional struggles are at the heart of inefficient policy choices and macroeconomic policy cycles, why do policy makers not design compensation schemes to neutralize political opposition? In the Alesina-Drazen (1991) model, for example, the distribution of the costs of stabilization is taken as exogenous. This is unsatisfactory because the design of the stabilization package can surely influence the distributional impacts. I will return to the issue of compensation toward the end of the paper. III. Crisis and Reform During the 1980s The 1980s experienced two events of lasting significance. First, much of the developing world became engulfed in a protracted debt crisis. Second, many countries began to shed their importsubstitution policies and endorsed market-oriented ones. The reforms did not come immediately after the crisis (Table 7). In fact, the typical pattern was for governments to respond to crisis by tightening their restrictions. Even Chile, which had already opened up during the 1970s under General Pinochet, initially chose to increase its (uniform) import tariff when the crisis hit. But after some delay, which differed across countries, countries in Latin America, Africa, and Asia jumped on the bandwagon of re18 John Pencavel has suggested that the policies of the Mexican petroleum labor union at PEMEX can be explained in such a framework. 19 The relevance of these considerations was suggested by Pencavel in his comments on this paper.
26
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 7 TIMING OF REFORMS AND INFLATION Annualized inflation rate prior to stabilization (%) 23,455 159 1,344 12,378 2,103 16
Stabilization Boliva Mexico Argentina Peru Brazil India 1985 1987–88 1991 1990 1994 1991
Structural Reform 1985 1985–1988 1987–1991 1990 1988–1990 1991
form. Governments endorsed reform with varying degrees of enthusisasm. The most enthusisastic reformers by far were in Latin America, where the Chilean example played a major role. (For more on Latin America, see Edwards 1994.) The developing countries were followed after 1989 by the socialist countries of Eastern Europe and the former Soviet Union, which embarked on their own transition to market economies. A. Does Crisis “Cause” Reform? It is natural to suppose that crisis and reform were somehow related. Indeed, if there is one single theme that runs through the length of the political economy literature it is the idea that crisis is the instigator of reform. This theme keeps reappearing in different guises. Ranis and Mahmood write that resistance [to] vested interests can be overcome only when the system has no other way of avoiding the required adjustment. In other words, it has to be “up against it,” without any easy alternative by which the day of reckoning can be delayed—either because a country’s own resources or the resources made available from the outside are considered insufficient to keep the system operating along the old rails. (1992, p. vi)
“When populist leaders in Argentina, Bolivia, Venezuela, Peru, and Brazil adopted nonpopulist policies,” claims
Bresser Pereira, “it was because the crisis in these countries was so deep that even the costs of sticking to populist policies became higher than the costs of adjustment” (1993, p. 57). According to Krueger, economic reforms were undertaken when “economic conditions deteriorated sufficiently so that there emerged a political imperative for better economic performance” (1993, p. 109). Moreover, the idea is not limited to developing countries. “The ‘golden rule’ of recent Spanish economic history,” writes Guillermo de la Dehesa “is: ‘Only when the level of reserves was sufficiently low and/or the current account balance was in large deficit have necessary economic adjustment and structural reform measures been taken’ . . .” (in Williamson 1994, p. 137; footnote omitted). Explaining why Australia undertook fewer reforms than New Zealand, Max Corden writes “[t]he reforms have been less dramatic than New Zealand’s because things never got so bad: inflation did not rise so high, and considerable trade liberalization had taken place since the 1950s” (in Williamson 1994, p. 112). The idea that governments have to be up against the wall before they act may be valid, but it is not entirely free of problems from an analytical standpoint. First, note that there is a strong element of tautology in the association of reform
Rodrik: Understanding Economic Policy Reform with crisis. Reform naturally becomes an issue only when current policies are perceived to be not working. A crisis is just an extreme instance of policy failure. That policy reform should follow crisis, then, is no more surprising than smoke following fire. 20 Furthermore, the hypothesis is virtually nonfalsifiable: if an economy in crisis has not yet reformed, the frequently proffered explanation is that the crisis has not yet become “severe enough.” Bresser Pereira, who was briefly Brazilian finance minister in 1987, argues that the reason his fiscal package was not supported by President Sarney was that the crisis was not yet perceived as serious (in Williamson 1994, pp. 333–54). True, perhaps, but not very helpful for explanation or prediction. What we surely need to understand is why South Korea’s politicians are ready to change course at the slightest hint of a crisis, while Brazil’s will bring their economy to the brink of hyperinflation several times before they tackle the problem. The political-economy literature recognizes this issue, but is largely silent on it. Some helpful hints come from the analytical literature mentioned previously. Alesina and Drazen (1991) show in their “war of attrition” model that the lower the “degree of cohesion” in society, which they model as the expected asymmetry in the burden of stabilization, the greater the delay before a stabilization takes place.21 One can loosely interpret this in the following terms. In societies where resources are evenly distributed, and the government has a good distributional track record, groups are less in20 Williamson and Haggard (in Williamson 1994, p. 564) mention several cases where reform was undertaken despite the absence of crisis (notably Australia, Colombia, and Portugal). 21 This is also possibly what Krueger has in mind when she argues that a “strong” government such as Korea’s will react quickly to crisis, while a “factional” one will delay (Krueger 1993, p. 126).
27
clined to believe that the burden of adjustment will be one-sided. This places the East Asian countries in an advantageous situation, once again, where crises are concerned. In addition, the emphasis on crisis has in itself little predictive content as to what form the response will take. Latin American countries eventually adjusted not only by balancing their fiscal accounts, but also by overhauling their trade and industrial policies and undertaking privatization on a major scale. As I have discussed above, these trade and industrial policies had little to do with instigating the crisis, and consequently their reform was not a logical necessity once decisive action was taken. In fact, some of the reforms in the area of trade liberalization almost surely complicated the macroeconomic stabilization effort.22 Therefore, the comprehensiveness of the reforms that the Latin American countries undertook still requires explanation. Part of the explanation has to do with the advice that these governments received. As mentioned previously, many professional economists themselves tended to lump together all “import substitution” policies and to hold them equally responsible for the crisis. That was certainly the approach taken by the World Bank, the most important conduit of economic ideas to developing-country policy makers. But how could these wide ranging
22 That is so for several reasons. The reduction of tariffs resulted in losses in fiscal revenue. Furthermore, trade liberalization aggravated the adverse effect of stabilization policies on aggregate demand and domestic output. Last but not least, it created an important conflict in exchange-rate management: trade liberalization required a “compensating” devaluation of the currency, which often could not be undertaken for fear of upsetting fragile price expectations. The recent Mexican debacle could possibly have been avoided, or at least its impact minimized, if the authorities had not removed controls on trade and capital flows so enthusiastically.
28
Journal of Economic Literature, Vol. XXXIV (March 1996) ciated with this reform can be expressed as 1/µεt, where µ is the share of affected imports in domestic consumption and ε is the absolute value of the import demand elasticity. With reasonable values of t = 0.5, µ = 0.2, and ε = 2, the formula yields a value of five for the PCBR. 23 That is, five pesos of income have to be reshuffled among groups in society to yield a single peso of efficiency gain! This is one way of capturing the political difficulty of achieving microeconomic reform in normal times. Now suppose that the economy gets caught up in a deep macroeconomic crisis, recovery from which would provide all-around gains. The idea here is that triple-digit inflation and the associated instability hurt pretty much all groups in society, although some may naturally protect themselves better than others. Suppose further that a reformist government packages the stabilization program needed to end the crisis with the price reform discussed above. Let γ stand for the percentage increase in income that accrues to all groups in society as a result of the stabilization. We can think of γ as an index of the depth of the macroeconomic crisis. Under reasonable assumptions on the value of γ (e.g., γ = 0.10) the PCBR plummets to well below unity (from five previously). The point is that the opportunity to do something that will benefit most everyone by a large margin—an opportunity that arises only when the economy is mismanaged terribly and falls into deep crisis—allows reformist policy makers to sneak in, along23 There are three groups that matter in this example, consumers of the importable, producers of the importable and the rest of the economy as affected by the government budget. Let the world price of the importable be p*, the domestic price p, and the quantities of consumption, production, and imports c, q, and m, respectively. Consumers gain -cdp, producers gain qdp, and the government budget changes by mdp + tp*dm. The efficiency gain is tp*dm. See Rodrik (1994b).
trade and industrial policy reforms be rendered palatable to the interest groups that had been their beneficiary for so long? How were they persuaded to go along? Here again it is plausible that the atmosphere of crisis played a role. Crisis enabled reformist governments to package fiscal reforms, which were absolutely crucial for the return of price stability, with trade and industrial policy reforms, which were viewed as desirable in the longer run but were incidental to the immediate crisis. In other words, policy makers acted as agenda setters: They presented domestic interests with a package of both macroeconomic and microeconomic reforms. Because high inflation and macroeconomic instability hurt pretty much everyone across the board, influential interest groups felt compelled to go along. They may have preferred to have only the macroeconomic component of the package, but that was not the choice that they confronted. The argument can be illustrated by using a simple heuristic concept that we may call the “political cost-benefit ratio” (PCBR, Rodrik 1994b). Assume that redistributing income is politically costly, and that this cost has to be traded off against the efficiency benefits of reform. Define the PCBR as the ratio of the total redistribution generated by reform to its efficiency benefits. This index answers the question: how many pesos are reshuffled among different groups for every peso of efficiency benefit? The PCBR takes values between zero and infinity. The higher the value of the index, the more difficult reform is likely to prove. Price reforms (such as trade liberalization and removal of subsidies) tend to generate a lot of redistribution relative to their efficiency benefits. Consider the removal of an import tariff, t. To a firstorder of approximation, the PCBR asso-
Rodrik: Understanding Economic Policy Reform side the stabilization, microeconomic, and structural reforms which have significant distributional implications and which would be difficult to implement under normal circumstances. B. Does Reform Have Short-Term Costs? The proposition that some of the major structural reforms of the 1980s and 1990s were adopted despite interest groups, rather than because of them, naturally raises the question of the sustainability of the reforms. Here we face two other conventional wisdoms of the political-economy literature: First, reforms become sustainable when they generate “winners” with a stake in their continuation. Second, and this one is the downer, reforms tend to make things worse before they make them better. For a proposition that is startlingly lacking in empirical support, the second piece of conventional wisdom is surprisingly strongly held. It permeates practically every discussion of the political economy of reform. “The results of many worthwhile reforms lie on a J-curve,” writes Piñera, “they tend to make things a good deal worse before they get better” (1994, p. 227). Bresser Pereira, Maravall, and Przeworski (1993, p. 2) state categorically: “the effect of economic reform on growth must be negative in the short run.” “A key political problem of sustaining support for reform programs,” argues Joan Nelson “is the long delay in reaping visible benefits for much of the population.” She continues: “Quite simply, in most cases there are not enough early winners to ensure the political sustainability of the program . . .” (in Williamson 1994, p. 476–77). The facts do not support such pessimism about long-delayed response. Once one makes allowance for the likelihood that the counterfactual—no reform— produces even worse results in the short
29
run, the consequences of reform actually look pretty good. This shows up in a number of different areas. With regard to disinflation, most of the recent cases of exchange-rate based stabilizations, such as those in Israel (1985), Mexico (1987), and Argentina (1991), have been accompanied by consumption booms, rather than recessions.24 With regard to broader structural adjustment policies, the best statistical evidence to date is that such policies tend to significantly increase, rather than decrease, growth of output within two or three years, even if not immediately (Corbo and Patricio Rojas 1992).25 Even in Eastern Europe, where there is genuine reason to believe that the transition may have short-term costs, the evidence indicates that reform reduces rather then intensifies the shortterm costs. Output has fallen the least in countries like Poland and the Czech Republic which have had the most extensive reforms, and the most in countries like the Ukraine which have had the least (Table 8). Furthermore, consumption statistics indicate that much of the measured contraction of GDP in transitional economies is illusory. (For the Polish case, see Andrew Berg 1993.) Of course, structural reforms (and possibly stabilization as well) may have sharp distributional consequences, as discussed previously. But this is not a matter of the short run versus the long run, as with the worries reflected in the quotes above. Saying that reform is diffi24 See for example Guillermo Calvo and Carlos Végh (1994). Dornbusch and Fischer (1993) argue that it may be unavoidable to incur output costs when disinflating from moderate inflation (40–80 percent per year), but their evidence on this is not so clear cut. 25 Corbo and Rojas use a control-group approach and focus on the 1985–88 performance of countries that received at least two Structural Adjustment Loans (SAL) from the World Bank starting in 1985 or before. They find that adjustment raised GDP growth rates by an average of two percentage points.
30
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 8 REFORM IN ECONOMIES IN TRANSITION Average annual real GDP change during first three years of transition (%)* Countries with Strong Reform Programs Poland −5.5 Czech Republic** −8.2 Hungary −6.2 Estonia −10.1 Countries with Weak Reform Programs Russia −14.7 Ukraine −14.4 Moldova −17.8
Annual inflation, 1994(%) 30 9 19 47 336 1000 245
Source: Calculated from IMF (Oct. 1994, Table 11). ∗ The first year of the reform program is taken to be 1990 for Poland and Hungary, and 1991 for others. ∗∗ The figures for Czechoslovakia are used for 1991 and 1992.
cult because some powerful groups will be inevitably made worse off—in the long run as well as the short run—is distinct from claiming that reform is difficult because the net benefits from the reform come too late for politicians to reap the gains. It is the latter that seems empirically problematic. None of this is to dismiss the obvious fact that reforms do arouse opposition, and that the opposition often tends to be strongest early on. The point is that we cannot really attribute these phenomena to presumed (rather than demonstrated) short-term costs of reform. Once irrationality and myopia are dismissed, the political economy of reform turns out once again to contain more puzzles waiting to be worked out. C. Does Foreign Aid Help Reform? If reform has short-term costs, as so many believe, then foreign aid should help reforms get launched (and sustained) by alleviating these costs. This indeed is the standard justification for the World Bank’s and the IMF’s “struc-
tural adjustment” lending as well as for official bilateral credits. But the logic of this statement is not unassailable, even if the premise is accepted. The reason is that external resources reduce the costs both of reform and of doing nothing— that is, avoiding reform. In addition, the prospect of aid can actually exacerbate the delay in stabilization, by inducing groups to postpone making sacrifices until aid actually materializes. 26 The effect on reform is consequently ambiguous. One of the strongest proponents of foreign aid has been Jeffrey Sachs, who has served as a high-profile advisor to reformist governments in Bolivia, Poland, and Russia among other places. In his contribution to Williamson (1994) entitled “Life in the Economic Emergency Room,” he complains that economists tend to neglect the role played by foreign assistance in most of the major postwar reforms. He provides capsule histo26 This argument is worked out in Alessandra Casella and Barry Eichengreen (1994), using the Alesina-Drazen model.
Rodrik: Understanding Economic Policy Reform ries of the economic reforms in postwar Germany, Japan, Bolivia, Mexico, Chile, Poland, Israel, and Turkey, stressing the critical contribution of aid in each case. Sachs is explicit that a government committed to reform is needed before aid can do any good. But once such a government is in place, the international community should be forthcoming. What aid can do, even if it is small in relation to the country’s problems “is help good governments to survive long enough to solve problems” (in Williamson 1994, p. 512). The problem is that aid can also help bad governments survive. For debating purposes, one can cite at least as many cases as Sachs does to demonstrate an association between plentiful aid and delayed reform. 27 As mentioned previously, one of the themes of the book by Ranis and Mahmood (1992) is that the availability of external resources tends to promote irresponsible policies. Scarcity of resources, on the other hand, is good for reform. (Note the obvious parallel to the crisis hypothesis.) One of the pieces of conventional wisdom about the Korean and Taiwanese reforms of the 1960s is that these reforms took place in large measure because U.S. aid, which had been plentiful during the 1950s, was coming to an end (see for example Ranis and Mahmood 1992, p. 139). Sachs would respond that donors must ensure the recipient governments will undertake the reforms before doling out the cash. More generally, aid must come with a heavy dose of conditionality. This is well recognized, and both the IMF and the World Bank make access to their resources conditional on good behavior
27 Some of Sachs’ case histories can also be challenged. It is true that aid played an important role in the Turkish reforms of the early 1980s, but it is also true that it delayed the fiscal day of reckoning, with ultimately debilitating consequences for the Turkish economy.
31
on the part of the borrowing governments. But conditionality is no panacea, as conditionality can last (at best) only as long as net transfers are positive. In the end, sovereign governments are just that: sovereign. There is a fairly large literature on conditionality and its effectiveness (see Manuel Guitian 1982; Paul Mosley 1987; Jacques Polak 1991; and V. Thomas et al. 1991). Some emerging conclusions are that conditionality is most effective when it focuses on a small range of quantifiable indicators and when governments are already committed to the reforms. However, it may not be easy to tell governments apart, and even governments that are genuinely reformist may be tempted by aid to deliver less than they promised. IV. In Search of a Manual for Reformist Politicians All of which raise the question, “what’s a poor reformist politican to do?” Recently Williamson had the interesting idea of gathering a group of high-ranking technocrats (“technopols” in Williamson’s jargon) to talk about their experiences, in the hope that some common lessons may emerge. The resulting book (Williamson 1994) opens with a list of hypotheses drawn from the literature about what makes reform feasible and successful, and which contributors were asked to examine from their own individual perspective. I have reproduced the list in Table 9. The list includes some items that I have already discussed, as well as many others. In the end, however, Williamson is forced to concede defeat. In his concluding chapter (written with Haggard), he writes “there are no fully robust empirical generalizations; in every case there is at least one partial counterexample. None of the 15 hypotheses investigated was either necessary or sufficient for
32
Journal of Economic Literature, Vol. XXXIV (March 1996)
TABLE 9 HYPOTHESES ABOUT REFORM
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
Policy reforms emerge in response to crisis Strong external support (aid) is an important condition for successful reform Authoritarian regimes are best at carrying out reform Policy reform is a right-wing-program Reformers enjoy a “honeymoon period” of support before opposition builds up Reforms are difficult to sustain unless the government has a solid base of legislative support A government may compensate for the lack of a strong base of support if the opposition is weak and fragmented Social consensus is a powerful factor impelling reform Visionary leadership is important A coherent and united economic team is important Successful reform requires economists in positions of political responsibility Successful reform requres a comprehensive program capable of rapid implementation Reformers should mask their intentions to the general public Reformers should make good use of the media Reform becomes easier if the losers are compensated Sustainability can be enhanced by accelerating the emergence of winners
Source: Williamson (1994)
successful reform” (p. 589). 28 The propositions that received strong support were “the need for a strong political base, for visionary leadership, and for a coherent economic team.” Disproved were the hypotheses that reform requires authoritarian regimes or that it is an exclusively right wing enterprise. But the volume itself is far from a failure. The exchanges among former politicians, technocrats, economists, and political scientists that it contains make for fascinating reading. A fundamental fault line that divides the contributors to this literature is the issue of how participatory reform politics ought to be. Most economists are on the side of speed, stealth, and consequently of reform from above. Vladimir Mau, a Russian economist, writes:
28 As indicated in Table 9, there were actually 16 hypotheses, so the number 15 must be a misprint.
Count Sergei Witte, a prominent reformer under the czars, used to say that there were two essential elements for radical reforms in Russia: absolute monarchy, because you need not pay attention to your critics if His Majesty supported you, and speed, because somebody might persuade the czar to change his mind before the reform could be made irreversible. (pp. 435–36)
Sachs cites approvingly the example of Jacques Necker (the French finance minister before the 1789 revolution) who maintained that flexibility or willingnesss to compromise, which might be harmless or even advantageous in other ministers, was an unforgivable failing in a finance minister!
Sachs goes on:
If reformers want free prices, they should not stand around and talk about it—they should do it, because everyone will be against freeing prices until it has been done, until it is an established fact. (pp. 509–10)
Sachs has plenty of first-hand experience in Bolivia, Poland, Russia, and elsewhere, so his views perhaps ought to
Rodrik: Understanding Economic Policy Reform count for a lot more than those of armchair economists. He is especially outspoken on the question of whether reformers should strive for a consensus with affected groups.29 “Many participants have suggested that reformers succeed by constructing a ‘social consensus’ in favor of reforms,” he writes. “This is mostly not the case. In deep crises, there is simply no consensus to build upon, only confusion, anxiety, and a cacophony of conflicting opinions.” He goes on to relate how the Polish “big bang” was enormously controversial at the time, and the desire to “return to Europe” did not translate into agreement on specific policies. “Consensus on many specifics (e.g., currency convertibility, price decontrol, budgetary discipline) has come to Poland, but only after three years of reform” (p. 505). In Sachs’ view, it is at best a waste of time to seek a broad coalition for reform because most people have no understanding of what is required: “While the history of marketbased reforms has repeatedly shown that free markets, open trade, and an economy fueled by private ownership are enormously powerful in stimulating rapid economic growth, the general public rarely knows it or believes it at the start” (p. 506). Neither is the problem limited to the person on the street: “Few Russians understand the source of Russia’s current inflation, least of all the [former] governor of the central bank . . .” (p. 507). The operational implication of all this is that reform needs a strong and autonomous executive, unhindered by the search for consensus and compromise. 30
29 Of course, there are very few things that Sachs is not outspoken about. His sharp critique of IMF policy in Russia in his after-dinner speech at the Williamson conference drew a defensive response from the IMF (printed in the Williamson volume), and (naturally) a rejoinder from Sachs. 30 In a personal communication, Sachs has clarified that his views should not be taken to imply
33
Sachs’ words have the advantage of expressing forthrightly what a lot of economists feel deep down but find politically incorrect to articulate. What some may find striking in these statements is the lack of faith in the common sense of ordinary people and in the efficacy of political institutions, especially in new democracies. That economists should hold to these views is not without irony. After all, homo economicus is supposed to be rational and forward-looking, and to process all the information that comes his way in the most efficient way. Sachs’ homo politicus is none of these things, or else is being held hostage to some grand coordination failure whose nature is unclear.31 This irony is seized upon by Przeworski, who provides a serious challenge to the economists’ preference for reform from above (in Bresser Pereira, Maravall, and Przeworski 1993). Focusing on the Polish case, he faults the storm tactics favored by Sachs and others for both weakening democratic political institutions and making errors in economic policy more likely. He and his two coauthors (Bresser Pereira and Maravall) summarize their argument in their concluding chapter thus: we find that subjecting the reform strategy to the competitive interplay of political forces is superior on three essential grounds: It
that he favors authoritarian forms of government over democracy. He believes, however, that democratically elected governments should have a fairly free hand in between elections. 31 One reader objected to this paragraph in the following terms: “The notion that the conventional economist may be a better judge of economic policy than the man on the street contains no more ‘irony’ than the notion that a physician may know better how to diagnose and cure my illness.” Most of us, however, take our physician’s advice. Were economists’ advice as well received by politicians, the puzzles on which this paper focuses would be far less salient.
34
Journal of Economic Literature, Vol. XXXIV (March 1996) a terrifying prospect for most Poles. 32 The main shortcoming of the Balcerowicz plan, therefore, was its failure to put in place adequate social protection to guard against the inevitable unemployment that would result. 33 Przeworski argues that this was a “technical” mistake, but one that resulted from not having listened to the people:
If the purpose of their architects was to make reforms politically palatable, the blueprint was not well designed. But the reason technocrats commit ‘technical’ errors is that they do not consult and concert with those who are affected by their blueprints. There is something paradoxical when believers in the informational efficiency of decentralized decisions fear them most . . . The main obstacle to reform is people. (p. 185)
32 That the decline in popularity was related to rising unemployment is plausible enough, but the statistical analysis that Przeworski undertakes to establish this relationship is unconvincing. The analysis consists of pairwise correlations between poll results and actual economic outcomes. The results of the exercise are unreliable for several reasons: there are few observations (a maximum of 20), there is obvious scope for omitted-variable bias, and a lot of the variables have strong time trends leading to spurious correlations. The strong negative correlation Przeworski obtains between the program’s popularity and unemployment is a statistical reflection of the fact that the first of these consistently went down, and the second up, after January 1990. A more telling piece of evidence comes from the experience of the Czech Republic (not discussed by Przeworski). Alone among reformist governments, Vaclav Klaus’ government has retained much of its popularity. Not coincidentally, the Czech Republic is also the only strongly reforming economy that has not (yet) experienced high unemployment. An alternative argument for the decline in the popularity of reforms is provided in Rodrik (forthcoming), based on changes in the preferences of state-sector workers over the course of the transition. 33 This point would be disputed by Balcerowicz and Sachs. According to Sachs (1994, p. 3): “Contrary to the impression of uncompromising laissez faire and harsh social policies, Poland’s social spending has risen very sharply since the beginning of reforms. Total budgetary spending on social programs (labor fund, pension fund, health care) has risen from 10 percent of GNP in 1990 to 21 percent of GNP in 1993, one of the highest proportions in Europe.”
improves policy, it builds support for the continuation of reforms, and it helps consolidate democratic institutions. We do not see a trade-off between public discussion and the soundness of economic plans. (p. 210)
These points are underscored in Przeworski’s interesting discussion of the Polish program. The program began with the famous “big bang” of January 1st, 1990. Leszek Balcerowicz, a contributor to the Williamson volume (and from whom more later), was its principal architect and Sachs its principal foreign advisor. Przeworski grants that the nearhyperinflation in which the Polish economy found itself by the Fall of 1989 required some fast action. But he faults the program for having had no social dimension: “I could find no references to building a welfare state or extending social protection in any formulation of the reform program,” he declares; “this was a pure trickle-down model of reforms” (p. 142). This may be bad enough for some, but the real clincher in Przeworski’s argument is that the absence of social policy ended up threatening the sustainability of the economic reforms. To demonstrate this, he analyzes at length the trend in public support for the Balcerowicz program. He finds that until May 1990 public support was high and steady, after which a significant decline in support occurred. Thereafter it remained steady again, until early 1991 when “the diagnosis of the current situation, optimism, willingness to bear sacrifices, and support for the Balcerowicz program all began to decline again” (pp. 158–59). Eventually, of course, support would fall so low that the former Communists would return to power in the September 1993 parliamentary elections. Przeworski attributes this slide in support primarily to the rise in unemployment, which he argues convincingly was
Rodrik: Understanding Economic Policy Reform
Moreover, Przeworski and his colleagues believe that the top-down style of these reforms tend to corrode nascent democratic institutions: the autocratic policy style characteristic of Washington-style reforms tends to undermine representative institutions, to personalize politics, and to generate a climate in which politics becomes reduced to fixes, to a search for redemption. Even if neo-liberal reform packages make good economics, they are likely to generate voodoo politics. (pp. 9–10)
35
Przeworski undertakes an interesting analysis of the trend in public confidence in various Polish institutions. He documents how confidence in representative institutions like political parties, the government, and the Sejm gradually eroded while confidence in the army and the police increased, to the point where “two years after the transition to democracy, Poland was a country in which the three institutions in which people had the most confidence were the army, the church, and the police” (p. 173). He attributes this to the “policy style” of the reforms— the introduction and continuation of a set of policies “by surprise and independently of public opinion and of representative organizations and institutions” 34 (p. 175). To paraphrase Przeworski only slightly, economic reform is consequently too important to leave to economists. The alternative to which Przeworski and his colleagues would subscribe is not entirely clear. In terms of actual policy content, their “social democratic” approach departs only mildly from the Washington consensus, in envisioning a somewhat more active role for the government in industrial policy and in coordinating private investment. 35 The rest
34 “The televised spectacle of the government ramming through the parliament a package of complicated legislation . . . took its toll” (p. 176). 35 To which Sachs would retort: “Governments that have reached hyperinflation cannot self-
of the program, including the emphasis on a “social safety net” (to use the Washington term), would not look out of place in a World Bank document. Przeworski mentions the Hungarian case approvingly, and implies that the more gradual approach adopted there was preferable because it paid greater attention to social issues. Yet, Hungarian unemployment has risen nearly as high as the Polish one, and by 1994 most economists would agree that the Polish economy had turned the corner while Hungary was still struggling. The cumulative output fall in Hungary has been larger, and the recovery smaller, than in Poland (see John Flemming 1995, Table 4). In any case, in 1994 former Communists were elected to power in Hungary as well. Perhaps an alternative model would have been the “social pacts” approach, used most conspicuously by post-Franco governments in Spain. This experience is described in Maravall’s contribution to the same volume (as well as by de la Dehesa in the Wiliamson volume). The Moncloa pacts in Spain were an effective device through which labor received expanded rights in return for wage moderation. But as Maravall makes clear, this experience is probably of little relevance to Eastern European countries, because there was still substantial room for expanded government in Southern Europe. Regarding the experience of Spain, Portugal, and Greece he writes:
Both public expenditure as a whole and social expenditure in particular increased very sharply from the mid 1970s under the new democratic regimes. Welfare systems and mechanisms for income distribution were extended and reorganized. The role of the state evidently, be expected to develop complex industrial policies or structural policies. After all, they aren’t even carrying out their most fundamental task” (in Williamson 1994, p. 510).
36
Journal of Economic Literature, Vol. XXXIV (March 1996) accompli than it would have been to build reforms gradually in the face of strong populist opposition in Parliament after the elections of October 1991. (pp. 172–73)
in the provision of health, education, and pensions was reinforced. Revenues from taxation increased, and fiscal systems were made more redistributive. (Bresser Pereira, Maravall, and Przeworski 1993, p. 101)
As he later puts it, “public expenditures could generally grow in spite of crisis” (Bresser Pereira, Maravall, and Przeworski 1993, p. 105). Leszek Balcerowicz, the former deputy prime minister after whom the Polish reform program was named, agrees with Przeworski’s facts, but not with the conclusions. There was no choice but to move quickly on both the macroeconomic and structural fronts, and public debate and deliberation would have wasted valuable time: “[T]he establishment of a social pact [as in Spain] in Poland in 1989 would simply have taken too much time during a crucial period.” Besides,
[w]hat would be the purpose of such a pact today? It could not afford to give compensation, because that would conflict with the fiscal program; the real compensation had consisted of the rapid elimination of shortages, which has been increasingly appreciated as time has passed. (in Williamson 1994, p. 218)
Moreover, it would have allowed the opponents of the program to dilute it. Like Sachs, Balcerowicz has little confidence in the political process in the short term, and complains that two election campaigns in the first two years of the program—presidential in 1990 and parliamentary in 1991—made his job extremely difficult. He sees the first few months of a honeymoon period in early 1990 as having provided an opportunity to ram through as many reforms as quickly as possible. “[T]he period of ‘extraordinary politics’ was short-lived and . . . one should use it to introduce tough economic measures.” Further:
[i]t was easier for the supporters of the market-oriented reforms to defend them as fait
One useful distinction in this debate, not often made, is between the initiation and consolidation of reform. Haggard argues that the initiation of reform requires independence or autonomy for the executive, while consolidation of reform necessitates “building of legislative and interest-group bases of support” (in Williamson 1994, p. 468). This argument is echoed by Joan Nelson, who says that it may be inevitable to be somewhat autocratic in the early stages of reform, but that in later stages one needs to generate support and consensus (pp. 472ff.).36 But what if the autocratic approach early on does irreparable damage to confidence in democratic institutions, as Przeworski claims it does? It is difficult not to feel sympathy for Przeworski’s yearning for a more democratic style of reform. Yet my suspicion is that most observers of the Polish reforms would agree with Balcerowicz and Sachs that speed (and stealth) was of the essence. While there have been some reversals in the reforms, even with these reversals the Polish economy is in far better shape today than it would have
36 Perhaps Mexico (prior to its most recent crisis) is a good example of the Haggard-Nelson strategy. José Córdoba relates in detail the dramatic changes in Mexican economic policy following the 1982 crisis. An astounding turnaround in the fiscal stance (of 10 percentage points of GDP!) was accomplished within two years, and in a fairly unilateral manner. But the government’s subsequent approach was more participatory (or at least, corporatist): “The Economic Solidarity Pact [of 1987] clearly established that inflation could not be reduced through government policies alone but required a social dialogue and active support by labor, business, and peasant organizations. That lesson profoundly impressed policymakers, and from then on much greater effort was made to build consensus and promote the direct involvement of those sectors in society that would be affected positively or negatively by the reforms” (in Williamson 1994, p. 254).
Rodrik: Understanding Economic Policy Reform been under a more gradual path. Balcerowicz is absolutely right that his reforms have created a new, largely irreversible reality that his successors dare not touch—a matter of importance when these successors are former Communists. This fascinating debate takes us back to the most fundamental issue in the political economy of policy reform, the one with which this essay opened: If the objective of reform is to make people better off, why does reform have to be shielded from the people? “[T]he sad fact that good things—such as democracy and market-oriented economic policy—do not always go together,” as Haggard puts it (in Williamson 1994, p. 467), suggests that there is a real puzzle here. Furthermore, if the problem with reform is that powerful groups are hurt by it, why can’t policy makers come up with compensation schemes that remove the hurdle? Myopia, to which many observers ultimately resort, appears to me to be as unsatisfactory an explanation here as in any area of conventional economics. As I pointed out above, the analytical literature has begun to address these puzzles. I close this section by discussing briefly two papers which are especially germane to the questions at hand. The first one (Raquel Fernandez and Rodrik 1991) provides an answer to why reforms that will benefit most everyone may still be rejected by a majority of the electorate, and why compensation may not remove distributional obstacles. The second (Mathias Dewatripont and Gerard Roland 1992) discusses optimal reform design under dynamic political constraints, restricting feasible policies to those that for example will be approved by unanimity. In Fernandez and Rodrik (1991) we ask whether a rational electorate would ever reject a reform which is known to
37
benefit a majority of the voters. We show that the answer is “yes.” We also show that political systems in general have a bias toward the status quo even when the status quo is inefficient and individuals are risk neutral. The key to the argument is uncertainty of a particular kind: the identity of many of the gainers (as well as losers) from reform cannot be determined ex ante. To see how the argument works, consider a democracy where a majority vote is needed before reform can be adopted. Let the economy have 100 voters and suppose that the reform in question will increase the incomes of 51 individuals by five zlotys each and decrease the incomes of the rest by one zloty each, leaving a net gain of (5 × 51) − (1 × 49) = 206 zlotys. In the absence of uncertainty, the majority of the population would vote in favor and the reform would be adopted. Assume that all these consequences of reform are common knowledge. Now suppose that while 49 individuals know for sure that they will gain, the remaining 51 are in the dark as to which among them will gain and which will lose. However, because aggregate consequences are common knowledge, individuals in the latter group know that two of them will eventually benefit while 49 will lose out. (Such uncertainty may arise, say, from incomplete information at the individual level about the skills needed to suceed in the post-reform environment.) This renders individuals in the second group identical ex ante, with an expected benefit from reform of [(5 × 2) − (1 × 49)]/51 = − 0.76 zloty each. Hence the individuals in the uncertain group will reject reform, blocking its adoption. The bottom line is that uncertainty about the consequences of reform at the level of the individual can prevent reform, even when it is recognized that reform will make a politically effective ma-
38
Journal of Economic Literature, Vol. XXXIV (March 1996) tial reform, which generates less of a budgetary burden. Dewatripont and Roland show that a dynamic version of this set-up produces an argument for gradualism. A partial reform early on would induce only the highest-ability group to leave; next period it becomes optimal to induce the next-highest group to leave; and so on, at much lower cost to the budget ultimately. The government’s assumed control over the agenda of policies under consideration allows it to dynamically relax the political constraints. A somewhat similar story, but using the FernandezRodrik (1991) model and based on the Chinese example, is provided by ShangJin Wei (1992), who shows how a sequence of reforms in different policy arenas may be politically sustainable where an economy-wide “big bang” would fail. V. Concluding Comments A political scientist or historian may well find much of the economics literature on the political economy of reform naive or simplistic. However, what is encouraging about this literature is that economists are now doing their politicaleconomy analysis explicitly, rather than implicitly as used to be the case. Most economists have now come to the realization that good economic advice requires an understanding of the political economy of the situation. The result has been a remarkable degree of collaboration between economists and political scientists, as well as more work on political economy by younger economists. Both of these are good news. The bad news is that the habit of attributing myopia or irrationality to political actors— whether explicitly or, more often, implicitly—persists. As this essay attests, there is by now a wealth of case-study material on the experiences with policy reform. This mate-
jority better off. 37 Moreover, the same kind of argument explains why compensation, or the promise thereof, is not always an effective device to remove the distributional obstacle to reform. In the above example, ex ante losers know that if reform is passed, there will be an ex post majority in support of its continuation—even in the absence of compensation. Therefore, a promise to compensate losers ex post is not going to be credible. This kind of reasoning may explain why many reforms that would have been popular ex post are passed up ex ante. It may also explain why reforms that are instituted by an authoritarian regime against prevailing political sentiment survive the return of democracy (e.g., Pinochet’s reforms in Chile). Inspired more directly by the Eastern European context, Dewatripont and Roland (1992) analyze the case of an agenda-setting government faced with the task of shrinking the size of the state sector subject to political constraints. Some of their analysis focuses on the case where a reform must be approved by unanimity to be passed. Efficiency calls for immediate shrinkage of the state sector (i.e., shock therapy), as productivity is much lower than in the private sector. But unanimity requires that statesector workers who are forced to exit be compensated. Suppose that state-sector workers are heterogeneous in terms of their earning power in the private sector (to which they will be displaced) and that this information is private. The budgetary cost of shock therapy will be large, as all exiting workers will have to be paid a common exit fee that is based on the income loss that would have been incurred by the least-capable worker. If, in turn, raising fiscal revenue is distortionary, shock therapy may be dominated by par37 The situation is worse when individuals are risk averse. But the argument does not rely on risk aversion.
Rodrik: Understanding Economic Policy Reform rial provides plenty to chew on for anyone interested in the interactions between economics and politics. The next step should be to integrate evidence from these case studies more systematically into the analytical work, and in turn to use the analytical models as a springboard for more rigorous case studies. I close by highlighting one of the many questions raised earlier as requiring more work. Because distributional issues are at the heart of the literature discussed here, we need more progress on understanding why institutions for compensating losers from reform are not more common. There are very few papers where the difficulties of compensation are made endogenous to the analytical framework.38 This makes the literature somewhat incomplete in its diagnosis of the issues. It also opens up a natural avenue for future research.
REFERENCES ALESINA , A LBERTO AND D RAZEN, A LLAN. “Why Are Stabilizations Delayed?” Amer. Econ. Rev., Dec. 1991, 81(5), pp. 1170–88. ALESINA , A LBERTO AND R ODRIK , D ANI. “Distributive Politics and Economic Growth,” Quart. J. Econ., May 1994, 109(2), pp. 465–90. ALESINA , ALBERTO AND T ABELLINI , G UIDO . “External Debt, Capital Flight and Political Risk,” J. Int. Econ., Nov. 1989, 27(3/4), pp. 199–220. AMSDEN, A LICE H. Asia’s next giant: South Korea and late industrialization. New York and Oxford: Oxford U. Press, 1989. BALASSA , BELA AND ASSOCIATES . The structure of protection in developing countries. Baltimore: Johns Hopkins U. Press, 1971. BARRO , ROBERT AND WOLF , H OLGER . “Data Appendix for Economic Growth in a Cross-Section of Countries.” Unpub. paper, 1989. BATES , R OBERT H. AND KRUEGER , A NNE O. Political and economic interactions in economic policy reform. Oxford: Blackwell, 1993. BERG , A NDREW . “Measurement and Mismeasure38 Dewatripont and Roland (1992) and Fernandez and Rodrik (1991) are exceptions. In Dewatripont and Roland, the difficulty is created by the unobservability of individual abilities. In Fernandez and Rodrik (1991) compensation turns out to be time-inconsistent when it is known that reform will garner an ex post majority even absent compensation.
39
ment of Economic Activity During Transition to the Market” in Eastern Europe in transition: From recession to growth. Eds.: M ARIO I. B LEJER ET AL . Washington, DC: World Bank, 1993, pp. 39–63. B HAGWATI , J AGDISH . Foreign trade regimes and economic development. Vol. 11. Anatomy and consequences of exchange control regimes. Lexington, MA: Ballinger, 1978. ———. India’s economy: The shackled giant. Oxford, Clarendon Press, 1993. B IRDSALL , N ANCY; R OSS , D AVID AND S ABOT, R ICHARD . “Inequality and Growth Reconsidered.” Unpub. paper, Feb. 1994. B OSWORTH , BARRY P.; DORNBUSCH , R UDIGER AND L ABÁN , R AÚL , eds. The Chilean economy: Policy lessons and challenges. Washington, DC: Brookings Institution, 1994. B RESSER PEREIRA , L UIZ CARLOS ; M ARAVALL , J OSÉ M ARIA AND PRZEWORSKI, A DAM . Economic reforms in new democracies: A socialdemocratic approach. Cambridge: Cambridge U. Press, 1993. CALVO , G UILLERMO A. AND VÉGH , C ARLOS A. “Inflation Stabilization and Nominal Anchors,” Contemp. Econ. Policy, Apr. 1994, 21(2), pp. 35–45. CASELLA , A LESANDRA AND EICHENGREEN, B ARRY. “Can Foreign Aid Accelerate Stabilization?” Unpub. paper, 1994. CLARKE , GEORGE R. G. “More Evidence on Income Distribution and Growth.” U. of Rochester, unpublished paper, 1993. COLLINS, S USAN M. AND P ARK , WON-A M . “External Debt and Macroeconomic Performance in South Korea,” in Developing country debt and economic performance, Vol 3. Eds.: JEFFREY D. S ACHS AND SUSAN M. COLLINS . Chicago: U. of Chicago Press, 1989, pp. 153–369. CORBO , VITTORIO AND FISCHER, S TANLEY. “Structural Adjustment, Stabilization and Policy Reform: Domestic and International Finance,” in Handbook of development economics, Vol. III. Eds.: J. R. B EHRMAN AND T. N. S RINIVASAN . Amsterdam: North-Holland, 1995. CORBO , V ITTORIO AND R OJAS , PATRICIO . “World Bank-Supported Adjustment Programs: Country Performance and Effectiveness,” in Adjustment lending revisited: Policies to restore growth. Eds.: VITTORIO CORBO , S TANLEY FISCHER , AND S TEVEN B. W EBB . Washington, DC: World Bank, 1992. pp. 23–36. CUKIERMAN , A LEX; EDWARDS , S EBASTIAN AND T ABELLINI , GUIDO . “Seignorage and Political Instability,” Amer. Econ. Rev., June 1992, 82(3), pp. 537–55. D E LONG , J. B RADFORD AND S UMMERS , L AWRENCE H. “Equipment Investment and Economic Growth,” Quart. J. Econ., May 1991, 106(2), pp. 445–502. D EWATRIPONT , M ATHIAS AND R OLAND , GERARD . “Economic Reform and Dynamic Political Constraints,” Rev. Econ. Stud., Oct. 1992, 59(4), pp. 703–30.
40
Journal of Economic Literature, Vol. XXXIV (March 1996)
I NTERNATIONAL M ONETARY F UND . World economic outlook. Washington, DC: The Fund, 1994. J ONES, L EROY P. AND SA KONG , I L. Government, business, and entrepreneurship in economic development: The Korean case. Cambridge, MA: Harvard U. Press, 1980. KIM , J ONG -I L AND LAU , LAWRENCE J. “The Source of Economic Growth of the Newly-Industrializing Countries on the Pacific Rim.” Unpub. paper. Stanford U., Dec. 1992. KRUEGER, A NNE O. Foreign trade regimes and economic development. Vol. 10. Liberalization attempts and consequences. Lexington, MA: Ballinger, 1978. ———. Political economy of policy reform in developing countries. Cambridge, MA and London: The MIT Press, 1993. L ABÁN, R AÚL AND S TURZENEGGER , F EDERICO . “Distributional Conflict, Financial Adaptation and Delayed Stabilization,” Economics and Politics, Nov. 1994a, 6(3), pp. 257–76. ———. “Fiscal Conservatism as a Response to the Debt Crisis,” J. Devel. Econ., Dec. 1994b, 45(2), pp. 305–24. L AGO , R ICARDO. “The Illusion of Pursuing Redistribution through Macropolicy: Peru’s Heterodox Experience, 1985–1990,” in R UDIGER D ORNBUSCH AND S EBASTIAN EDWARDS 1991, pp. 263–323. L IN, CHING - YUAN. Industrialization in Taiwan, 1946–72: Trade and import-substitution policies for developing countries. New York: Praeger, 1973. L ITTLE , I AN M. D.; S CITOVSKY, T IBOR AND S COTT, M AURICE . Industry and trade in some developing countries. London: Oxford U. Press, 1970. L ITTLE , I AN M. D. ET AL . Boom, crisis, and adjustment: The macroeconomic experience of developing countries. New York: Oxford U. Press for the World Bank, 1993. MONDINO, G UILLERMO ; S TURZENEGGER , FEDERICO AND T OMMASI , M ARIANO . “Recurrent High Inflation and Stabilization: A Dynamic Game.” Unpub. paper, 1992. M OSLEY, PAUL . Conditionality as bargaining process: Structural-adjustment lending, 1980–86. Princeton, NJ: Princeton Essays in International Finance No. 168, 1987. Ö ZLER , SULE AND T ABELLINI , G UIDO . “External Debt and Political Instability.” National Bureau of Economic Research Working Paper, 1991. PANAGARIYA , ARVIND . “India: A New Tiger on the Block?” J. International Affairs, Summer 1994, 48(1), pp. 193–221. Penn World Table 5.5. A LAN H ESTON ET AL . Distributed by the National Bureau of Economic Research. Cambridge, MA, 1993. PEROTTI , R OBERTO . “Increasing Returns to Scale, Politics and the Timing of Stabilization.” Columbia U., unpublished paper, 1992a.
D ÍAZ-ALEJANDRO , CARLOS F. “Trade Policies and Economic Development,” in International trade and finance: Frontiers for research. Ed.: PETER B. KENEN . New York: Cambridge U. Press, 1975, pp. 93–150. D OLLAR , D AVID . “Outward-Oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976–1985,” Econ. Devel. Cult. Change, Apr. 1992, pp. 523–44. D ORNBUSCH , R UDIGER AND E DWARDS , SEBASTIAN , eds. The macroeconomics of populism in Latin America. Chicago and London: U. of Chicago Press, 1991. D ORNBUSCH , R UDIGER AND F ISCHER, S TANLEY. “Moderate Inflation,” World Bank Econ. Rev., Jan. 1993, 7(1), pp. 1–44. EASTERLY, WILLIAM . “How Much Do Distortions Affect Growth?” J. Monet. Econ., 1993, 32(2), pp. 187–212. EASTERLY, WILLIAM ; R ODRÍGUEZ , CARLOS A LFREDO AND S CHMIDT -H EBBEL , K LAUS . Public sector deficits and macroeconomic performance. New York: Oxford U. Press, 1994. EDWARDS, SEBASTIAN. “Trade and Industrial Policy Reform in Latin America.” NBER Working Paper No. 4772, June 1994. FERNANDEZ , R AQUEL AND R ODRIK , D ANI . “Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty,” Amer. Econ. Rev., Dec. 1991, 81(5), pp. 1146– 55. FISHLOW , ALBERT ET AL . Miracle or design? Lessons from East Asia. Washington, DC: Overseas Development Council, 1994. FLEMMING , JOHN. “Equitable Economic Transformation.” Paper presented to the Conference on Income Distribution and Sustainable Growth, IMF, Washington, DC, June 1995. GRINDLE , MERILEE S. AND T HOMAS , J OHN W. Public choices and policy change: The political economy of reform in developing countries. Baltimore and London: Johns Hopkins U. Press, 1991. GUITIAN, M ANUEL. “Economic Management and International Monetary Fund Conditionality,” in Adjustment and financing in the developing world. Ed.: T ONY KILLICK . Washington, DC: International Monetary Fund, 1982, pp. 73– 104. H AGGARD , STEPHAN AND KAUFMAN , R OBERT R., eds. The politics of economic adjustment. Princeton, NJ: Princeton U. Press, 1992. H AGGARD , STEPHAN AND WEBB, S TEVEN , eds. Voting for reform: Democracy, political liberalization, and economic adjustment. New York: Oxford U. Press, 1994. H ARBERGER, A RNOLD C. “Secrets of Success: A Handful of Heroes,” Amer. Econ. Rev., May 1993, 83(2), pp. 343–50. H ESTON , A LAN AND SUMMERS , R OBERT . “A New Set of International Comparisons of Real Product and Price Levels: Estimates for 130 Countries, 1950–1985,” Rev. Income Wealth, 1988, 34, pp. 1–25.
Rodrik: Understanding Economic Policy Reform
———. “Income Distribution, Politics, and Growth,” Amer. Econ. Rev., May 1992b, 80(2), pp. 311–16. PERSSON, T ORSTEN AND T ABELLINI , GUIDO . “Is Inequality Harmful for Growth?” Amer. Econ. Rev., June 1994, 84(3), pp. 600–21. PIÑERA , J OSÉ . “Chile,” in J OHN WILLIAMSON, ed. 1994. POLAK , JACQUES J. The changing nature of IMF conditionality. Princeton, NJ: Princeton Essays in International Finance No. 184, 1991. PRZEWORSKI , ADAM . Democracy and the market: Political and economic reforms in Eastern Europe and Latin America. Cambridge: Cambridge U. Press, 1991. R ANIS , GUSTAV AND M AHMOOD , S YED AKHTAR. The political economy of development policy change. Cambridge, MA and Oxford: Basil Blackwell, 1992. R ODRIK , D ANI . “King Kong Meets Godzilla: The World Bank and The East Asian Miracle,” (1994a) in ALBERT FISHLOW ET AL. 1994. ———. “The Rush to Free Trade in the Developing World: Why So Late? Why Now? Will It Last?” (1994b) in S TEPHAN H AGGARD AND S TEVEN B. WEBB, eds. 1994. ———. “Getting Interventions Right: How South Korea and Taiwan Grew Rich,” in Econ. Policy, Apr. 1995a, 20, pp. 55–107. ———. “Trade and Industrial Policy Reform,” in Handbook of development economics. Vol. IIIB. Eds.: J. R. B EHRMAN AND T. N. S RINIVASAN. Amsterdam: North-Holland, 1995b. ———. “The Dynamics of Political Support for Reform in Economies in Transition,” J. Japa-
41
nese Int. Econ., Special Issue on Transition, forthcoming. S ACHS , J EFFREY. “The Political Challenge of Economic Transition: The Case of Poland.” Unpub. paper. Jan. 1994. T AYLOR , L ANCE , ed. The rocky road to reform. Cambridge, MA: The MIT Press, 1993. T HOMAS , V INOD ET AL. Restructuring economies in distress: Policy reform and the World Bank. Oxford and New York, Oxford U. Press, 1991. T OMMASI , M ARIANO AND V ELASCO , A NDRÉS . “Where Are We in the Political Economy of Reform?” Unpub. paper, May 1995. V ELASCO , A NDRES . “Are Balance-of-Payments Crises Rational?” New York U., unpub. paper, 1994. WADE , R OBERT . Governing the market: Economic theory and the role of government in East Asian industrialization. Princeton, NJ: Princeton U. Press, 1990. WEI , S HANG -J IN. “Gradualism versus Big Bang: Speed and Sustainability of Reforms.” Unpub. paper. Harvard U., Dec. 1992. WILLIAMSON , J OHN, ed. The political economy of policy reform. Washington, DC: Institute for International Economics, 1994. WORLD BANK . The East Asian miracle. Washington, DC: World Bank, 1993. Y OUNG , A LWYN. “Lessons from the East Asian NICs: A Contrarian View.” NBER Working paper No. 4482. 1993. ———. “The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth Experience.” Unpub. paper. Feb., 1994.
You May Also Find These Documents Helpful
-
Cited: Mohrig, Jerry R., Christina N. Hammond, Paul F. Schatz, and Terence C. Morrill.Modern Projcts…
- 1114 Words
- 12 Pages
Better Essays -
This paper contains THIRTY-EIGHT questions in FIVE sections: Sections A, B, C, D and E. Answer THIRTY-FOUR questions only…
- 448 Words
- 2 Pages
Satisfactory Essays -
The search for stable foundations of policy in the face of drastic and revolutionary change poses a problem of almost excruciating difficulty for all nations...…
- 4652 Words
- 19 Pages
Best Essays -
Evaluate possible economic policies, other than increasing the age limit, that a government might use to reduce significantly the consumption of alcoholic drinks.…
- 1305 Words
- 4 Pages
Better Essays -
The Constitution is an outline framework of fundamental laws for governing the United States and guaranteeing the basic rights of its citizens. The founding fathers wanted a democratic and balanced nation under control without too much interference from the government. In the Constitution, two of the six principles that we have, is a limited government, to prevent tyranny of the despot, as well as checks and balances. Limitation in the government is where the governmental power is restricted by law. No one has ultimate power so that we can all live in a free and equal society.…
- 1070 Words
- 5 Pages
Good Essays -
The following three paragraphs have been taken from a paper by an EKU student Sue Mattingly.…
- 466 Words
- 2 Pages
Good Essays -
In this part of the paper, you respond to and refute the counter-argument based on evidence discussed in Part Three. This response will draw on the earlier argument in Part Two to show that the original thesis can be defended against this objection.…
- 472 Words
- 2 Pages
Good Essays -
The paper cites a minimum of three sources, one from the University Library and another from the course textbooks or Electronic Reserve Readings, to support the paper’s position.…
- 298 Words
- 3 Pages
Satisfactory Essays -
Jared Doerr 11/17/14 Arguments and Reasoning Gladiator Events: Sport or Not? Are gladiator events of ancient Rome sport? Gladiator events were most certainly a violent form of entertainment and because of the violence I would think we would naturally feel that it is not a sport; however, before we can clarify whether or not gladiator events are sport, we first have to define what a “sport” is considered to be. I have broken down “sport” to contain three elements: some aspect of competiveness (a desire to win), some kind of physical exertion (strain on the body) and also some kind of mental exertion (problem solving skills). Based on these three elements, gladiator events in ancient Rome are most certainly sport.…
- 771 Words
- 4 Pages
Better Essays -
References: Goldman, H., Morrissey, J., Ridgley, S., Frank, R., Newman, S., & Kennedy, C. (1992). Lessons…
- 1683 Words
- 7 Pages
Powerful Essays -
“The wealth qualification for the aspiring politician is taken for granted; a governmental system dedicated to the promotion of personal wealth in a few selected areas will never act for the public good. The time has come, nevertheless, for citizens to demand, from their government, a measure of socialization—the provision of amenities for the many, of which adequate state pensions and sickness benefits, as well as nationalized transport, should be priorities. (302)”…
- 321 Words
- 2 Pages
Good Essays -
| The greatest strength of this paper is all the information provided. The statistics give the reader a chance to actually absorb numbers.…
- 2034 Words
- 9 Pages
Better Essays -
Note: Michael Ling is the sole author of this document. You’re welcomed to use its contents but, as a courtesy, please quote the source of this paper http//www.michaelling.net/…
- 2813 Words
- 12 Pages
Powerful Essays -
This paper may be reproduced in whole or in part with appropriate recognition to the author,…
- 28457 Words
- 131 Pages
Good Essays -
Abstract: Is there a role for government in the economy? Yes, says Heritage analyst Karen Campbell—but the government must focus on maintaining economic stability. Fiscal responsibility is an important part of that stability. Government debt can quickly become a burden on the economy and weaken its foundations. Sound macroeconomic policies enhance the credibility of the government and strengthen the political institutions. This credibility is vital for economic stability and Americans’ long-term investment decisions that allow the U.S. economy to flourish. In order to restore economic stability, policymakers must focus on restoring the institutional role of governing. Government can provide a stable environment for economic growth when it can be depended upon to maintain the stability of the currency, enforce and defend property rights, and provide oversight that assures private citizens that their transaction partners in the marketplace are held accountable.1 This will allow market participants to begin putting their resources back to work in the areas where they are most beneficial.2 After decades of lecturing developing countries on how to emerge from economic crisis and stimulate economic growth through sound government policies, U.S. policymakers and some economists are throwing out all their advice during the first major crisis test. This is particularly true when it comes to advice on accumulating more and more debt.…
- 3462 Words
- 14 Pages
Best Essays